literature review venn diagram

Conducting Literature Reviews: Developing Search Strategies

• Literature Review Basics
• Types of Literature Reviews
• Clarifying the Goal
• Choosing Databases and Literature Sources

Developing Search Strategies

• Choosing What Literature to Use
• Applying Screening Techniques
• Doing the Review
• Documenting and Writing Up the Review

When conducting a literature search, it is important to develop a strategy so that your literature review is comprehensive and complete. Therefore it helps to develop a search strategy. The following are tips that can help you as you develop your search strategy. 

• Using a Framework
• Identifying Search Terms
• Using Boolean Operators
• Using Index Terms/Thesauruses
• Incorporating Search Filters
• Determining Your Results Format
• Documenting Your Work

A search strategy/framework is important to develop to help you identify how the components of your research question can turn into search terms. It may be helpful to use a Venn diagram to envision what you need to search for. While the ultimate goal is to come up with research in the center of the Venn diagram, other studies that tie into the Venn diagram may help build your literature review down towards the research in the center. 

Coming up with search terms is important so that you can conduct your search effectively. It is a good idea to brainstorm possible search terms, including synonyms for the words, and also to conduct preliminary research so as to obtain some knowledge about what terms other researchers use when they are discussing a topic. 

Boolean operators such as AND, OR, and NOT can also be used to find terms. AND will require both terms or phrases to be in the citation, OR will require either one citation or the other, and NOT excludes terms. Boolean operators also include the truncation symbol * which allow for varying endings. Many databases also let you search for phrases. 

Using thesauruses or index terms help to identify terms via controlled vocabulary, so that you can identify how the database is calling a particular subject. Many databases also index and document citations with this vocabulary, therefore facilitating the finding of articles. When developing a literature review, it can therefore be helpful to use thesauruses and indexing terms. 

Filters or limiters can be used to further limit results, and are useful after you retrieve a list of results. Using filters, you can select different aspects such as, for example, articles within the last five years, methodology, language, and more. 

It is also important to identify the format of the works in a database. For example, some databases have only the citation and abstracts, some only the citation, and others have the full text. Knowing what to expect will help you, but do not just search the citations for a term. 

While doing all of the previous steps is important for conducting a good search, it is also important to document what was done. This allows you to remain organized and make revisions as necessary, and will help you document your methods section and help the process to be replicated. 

• << Previous: Choosing Databases and Literature Sources
• Next: Choosing What Literature to Use >>
• Last Updated: Jan 28, 2022 4:19 PM
• URL: https://library.divinemercy.edu/lit-review

• Research Guides

Literature Review: A Self-Guided Tutorial

Using concept maps.

• Literature Reviews: A Recap
• Peer Review
• Reading the Literature
• Developing Research Questions
• Considering Strong Opinions
• 2. Review discipline styles
• Super Searching
• Finding the Full Text
• Citation Searching This link opens in a new window
• When to stop searching
• Citation Management
• Annotating Articles Tip
• 5. Critically analyze and evaluate
• How to Review the Literature
• Using a Synthesis Matrix
• 7. Write literature review

Concept maps or mind maps visually represent relationships of different concepts. In research, they can help you make connections between ideas. You can use them as you are formulating your research question, as you are reading a complex text, and when you are creating a literature review. See the video and examples below.

How to Create a Concept Map

Credit: Penn State Libraries ( CC-BY ) Run Time: 3:13

• Bubbl.us Free version allows 3 mind maps, image export, and sharing.
• MindMeister Free version allows 3 mind maps, sharing, collaborating, and importing. No image-based exporting.

Mind Map of a Text Example

Credit: Austin Kleon. A map I drew of John Berger’s Ways of Seeing in 2008. Tumblr post. April 14, 2016. http://tumblr.austinkleon.com/post/142802684061#notes

Literature Review Mind Map Example

This example shows the different aspects of the author's literature review with citations to scholars who have written about those aspects.

Credit: Clancy Ratliff, Dissertation: Literature Review. Culturecat: Rhetoric and Feminism [blog]. 2 October 2005. http://culturecat.net/node/955 .

• << Previous: Reading the Literature
• Next: 1. Identify the question >>
• Last Updated: Feb 22, 2024 10:53 AM
• URL: https://libguides.williams.edu/literature-review

The Literature Review: 5. Organizing the Literature Review

• 1. Introduction
• 2. Why Do a Literature Review?
• 3. Methods for Searching the Literature
• 4. Analysing the Literature
• 5. Organizing the Literature Review
• 6. Writing the Review

1. Organizing Principles

A literature review is a piece of discursive prose, not a list describing or summarizing one piece of literature after another. It should have a single organizing principle:

• Thematic - organize around a topic or issue
• Chronological - sections for each vital time period
• Methodological - focus on the methods used by the researchers/writers

4. Selected Online Resources

• Literature Review in Education & Behavioral Sciences This is an interactive tutorial from Adelphi University Libraries on how to conduct a literature review in education and the behavioural sciences using library databases
• Writing Literature Reviews This tutorial is from the Writing section of Monash University's Language and Learning Online site
• The Literature Review: A Few Tips on Conducting It This guide is from the Health Services Writing Centre at the University of Toronto
• Learn How to Write a Review of the Literature This guide is part of the Writer's Handbook provided by the Writing Center at the University of Wisconsin-Madison

2. Structure of the Literature Review

Although your literature review will rely heavily on the sources you read for its information, you should dictate the structure of the review. It is important that the concepts are presented in an order that makes sense of the context of your research project.

There may be clear divisions on the sets of ideas you want to discuss, in which case your structure may be fairly clear. This is an ideal situation. In most cases, there will be several different possible structures for your review.

Similarly to the structure of the research report itself, the literature review consists of:

• Introduction

Introduction - profile of the study

• Define or identify the general topic to provide the context for reviewing the literature
• Outline why the topic is important
• Identify overall trends in what has been published about the topic
• Identify conflicts in theory, methodology, evidence, and conclusions
• Identify gaps in research and scholarlship
• Explain the criteria to be used in analysing and comparing the literature
• Describe the organization of the review (the sequence)
• If necessary, state why certain literature is or is not included (scope)

Body - summative, comparative, and evaluative discussion of literature reviewed

For a thematic review:

• organize the review into paragraphs that present themes and identify trends relevant to your topic
• each paragraph should deal with a different theme - you need to synthesize several of your readings into each paragraph in such a way that there is a clear connection between the sources
• don't try to list all the materials you have identified in your literature search

From each of the section summaries:

• summarize the main agreements and disagreements in the literature
• summarize the general conclusions that have been drawn
• establish where your own research fits in the context of the existing literature

5. A Final Checklist

• Have you indicated the purpose of the review?
• Have you emphasized recent developments?
• Is there a logic to the way you organized the material?
• Does the amount of detail included on an issue relate to its importance?
• Have you been sufficiently critical of design and methodological issues?
• Have you indicated when results were conflicting or inconclusive and discussed possible reasons?
• Has your summary of the current literature contributed to the reader's understanding of the problems?

3. Tips on Structure

A common error in literature reviews is for writers to present material from one author, followed by information from another, then another.... The way in which you group authors and link ideas will help avoid this problem. To group authors who draw similar conclusions, you can use linking words such as:

• additionally

When authors disagree, linking words that indicate contrast will show how you have analysed their work. Words such as:

• on the other hand
• nonetheless

will indicate to your reader how you have analysed the material. At other times, you may want to qualify an author's work (using such words as specifically, usually, or generally ) or use an example ( thus, namely, to illustrate ). In this way you ensure that you are synthesizing the material, not just describing the work already carried out in your field.

Another major problem is that literature reviews are often written as if they stand alone, without links to the rest of the paper. There needs to be a clear relationship between the literature review and the methodology to follow.

• << Previous: 4. Analysing the Literature
• Next: 6. Writing the Review >>
• Last Updated: Feb 8, 2022 5:25 PM
• URL: https://libguides.uwi.edu/litreviewsoe

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List

A consistent and general modified Venn diagram approach that provides insights into regression analysis

Robert m. o’brien.

Department of Sociology, University of Oregon, Eugene, Oregon, United States of America

Associated Data

All relevant data are within the paper.

Venn diagrams are used to provide an intuitive understanding of multiple regression analysis and these diagrams work well with two variables. The area of overlap of the two variables has a one-to-one relationship to the squared correlation between them. This approach breaks down, however, with three-variables. Making the overlap between the pairs of variables consistent with their squared bivariate correlations often results in the overlap of two of these variables with the third variable that is not the same as the variance of the third variable accounted for by the other two variables. I introduce a modified Venn diagram approach that examines the relationships in multiple regression by using only two circles at a time, provides a new and consistent reason why the circles need to be of the same size, and designates a “target variable” whose overlap with the other circle corresponds to the variance accounted for by the other variable or variables. This approach allows the visualization of the components involved in multiple regression coefficients, their standard errors, and the F -test and t -test associated with these coefficients as well as other statistics commonly reported in the output of multiple regression programs.

Introduction

The use of Venn diagrams in statistics can provide a way to make concepts such as variance accounted for in the dependent variable, multiple regression coefficients, the effects of multicollinearity between the independent variables on standard errors, and associated significance tests more intuitive to students and professionals [ 1 – 6 ]. The traditional Venn diagram approach does this by making an analogy between the proportion of area of overlap between circles that represent two variables and the proportion of variance accounted for. This works clearly and simply for bivariate regression, but it only works sometimes in situations with two or more independent variables. This inconsistency creates problems for using the traditional Venn diagram approach to represent multiple regression problems where there are two or more independent variables.

This inconsistency limits the usefulness of the traditional Venn diagram approach. In this paper a modified Venn diagram approach is outlined that addresses the problems with the traditional approach when there are two or more independent variables. I carefully lay out the modified Venn diagram approach and show how these diagrams relate to the most commonly reported statistics associated with multiple regression. I show how the diagrams relate to the formulas for these statistics.

The modified Venn diagram approach that I propose allows for an improved understanding of regression analysis. This means giving up, however, the representation of the individual effects of the k independent variables in a multiple regression analysis. Dispensing with diagrams having three or more circles allows a consistent representation of the standard features of regression analysis diagrammatically. I concentrate on the standard output from multiple regression analysis programs, including; R y 2 , the F -test for the significance of R y 2 , the regression coefficients and their t -test for statistical significance, the standard errors of the regression coefficients, variance inflations factors, and the Analysis of Variance of regression table. More can be done with the proposed system, but I leave that for others to explore. Although we give up something in the process of not using a single Venn circle for each of the independent variables; we gain much by doing so. It allows us to leave behind the inconsistent representations inherent in the traditional use of Venn diagrams in statistics.

Two examples from the traditional approach

The traditional Venn diagram approach works in the bivariate case: one independent variable and one dependent variable. For convenience of representation, each variable has the same variance and same area. This is accomplished by using standardized variables, so that each variable has a variance of one.

Fig 1 shows the simple bivariate situation of one independent variable ( x ) and a dependent variable ( y ). The correlation between the two variables is .50 or −.50, one cannot determine which from the diagram (the diagram works in both situations). The squared correlation coefficient is represented by the area of overlap between the two variables, which is .25 ( = r x y 2 = . 5 2 = − . 5 2 ) . The variance that is not accounted for by x in the dependent variable is .75 ( = 1 − r x y 2 ) . The squared standardize regression coefficient of y regressed on x is the area in y uniquely accounted for by x divided by the variance in x that is not associated with the other independent variables in the model. Since x is the only independent variable in the model, the area in y that is associated uniquely with x is .25. There is no other independent variable in the model; thus, the area of x not associated with the other independent variables is 1.00. Therefore, the standardized regression coefficient squared is .25 ( β y x 2 = . 25 / 1.00 ) and | β yx | = |.5|. To obtain the absolute value of the unstandardized regression coefficient, | b yx |, we multiply the absolute value of the standardized coefficient by ( sd y / sd x ), where sd y is the standard deviation of y and sd x is the standard deviation of x . I could go on (as we will see in the next sections), but these are the basic statistics typically discussed in the bivariate regression representations that use Venn diagrams.

This approach works well in the two-variable case (two Venn circles). Note that the diagram represents squared quantities such as squared correlations, the proportion of the variance (a squared quantity) that is accounted for by the independent variable, the proportion of the variance that is not accounted for by the independent variable, and the squared standardized regression coefficient.

Fig 2 depicts a situation in which r 12 = ±.70, r y 1 = ±.40, and r y 2 = .00. There is no problem in presenting the bivariate overlaps with the Venn diagram in Fig 2 . There is a fundamental problem, however, with the diagram in Fig 2 that has to do with the overlap of x 1 and x 2 with y . The total overlap of the two independent variables with the dependent variable in that diagram is .160 while R y ∙ 12 2 = .314. The combined variance in y accounted for by x 1 and x 2 is .314 not .160. This is easily verified using the equation for R y ∙ 12 2 found in many intermediate statistics texts:

In our case, because r y 2 = 0, R y ∙ 12 2 = ( . 16 + 0 − 0 ) / ( 1 − . 49 ) = . 314 . The variance not accounted for by the two independent variables is .686 and not .840 as implied by the tradition Venn diagram approach. This problem with the variance accounted for in y by the two independent variables is a case of “suppression” [ 7 – 9 ], which is not handled well by Venn diagrams. In the two independent variable case, suppression occurs when the variance accounted for in the dependent variable increases when one of the two independent variables is controlled for by the other. In the present case, R y ∙ 12 2 > r y 1 2 + r y 2 2 .

The variance shared by the two independent variables is .49 as implied by the diagram, but this shared variance for the independent variables can be misleading in the case of three independent variables. With three independent variables; for example, suppression can occur among these variables and make the total overlap of the independent variables based on their bivariate relationships to each other misleading.

Two independent and one dependent variable using a modified Venn diagram approach

In Fig 3 , I present a modified Venn diagram with the same data used in Fig 2 , but this diagram is consistent in terms of overlaps and the variance accounted for in the “target variables.” I first show how this modified Venn diagram approach can be used to illustrate the regression components of most interest in the two independent variable case. In the following section, I generalize this approach to k independent variables. Achieving consistency between the proportions of variance accounted for and proportion of area overlapped requires modifications in the traditional approach to using Venn diagrams for statistical interpretations. These modifications include: (1) The number of circles in the diagram considered at any one time is no greater than two (see [ 2 ]) for an excellent example of what can be accomplished by concentrating on just two circles in a Venn diagram]. (2) Each of the circles has the same size, but this is not because they are standardized to have variances of 1.00. They are of the same size to represent that each variable or combinations of variables can account for a proportion of variance in the “target” variable of from 0 to 1.00. (3) In my two circle diagrams, one of the circles represents one or more variables and the other circle is considered the target variable: the target variable is the one for which the proportion of the area overlapped by the other circle is proportional to the variance accounted for in that variable. (4) We assign the overlap of the dependent variable with the combined independent variables into a portion that is uniquely associated with the independent variable of interest (the target independent variable) and a portion that is not uniquely associated with the target variable.

Circles of the same size allow for the proportion of the area overlapped in the target variable to correspond to the proportion of the variance accounted for. This innovation in the rationale for using circles of the same size is necessary because the circle representing the combined effects of x 1 and x 2 , for example, is not a standardized variable with a variance of one. It is of the same size as the target variable so that it can overlap in area with the target variable from 0 to 1.00, which corresponds to the variance in the target variable that can be statistically associated with one or more variables. The modified Venn diagram approach provides a diagrammatic visualization of the statistical components that provide all of the standard output for basic multiple regression analyses. I focus on the basic regression analysis components to illustrate the modified Venn diagram approach.

Fig 3 contains two panels or rows of diagrams. The top panel displays the overlap of the two independent variables, the target variable (the one for which we compute the regression coefficient and associated statistics: x 2 ). This diagram is easily handled by the traditional Venn diagram approach or this modified approach. The overlap of the target variable with x 1 is .49 in the diagram and .51 of the target variables’ variance is independent of the other independent variable or not accounted for by the other independent variable in the model. We have labelled this independent or unaccounted for variance in the target independent variable as “ area a ” in the diagram.

The second panel of Fig 3 , on the left-hand side, displays the combined overlap of x 1 and x 2 with y : the cross hatched area is the proportion of the variance in y accounted for by x 1 and x 2 . This cross hatched area is .314 of the area of y (here y is the target variable and in this case the dependent variable); this area is labeled “ area b .” This diagram represents the area of overlap, corresponding to R y ∙ 12 2 , for the two independent variables correctly (which was not the case in Fig 2 ). The black section of y is the proportion of the variance of y that is not accounted for by the two independent variables in the model and is labelled “ area c .” The strategy here is to show the total overlap using a single circle to represent the effects of both x 1 and x 2 .

The diagram, to the right in this panel shows the amount of the variance in y explained by x 1 and x 2 again, but this time it is broken into a part that is associated with x 2 (the cross-hatched area: “ area d ”) after allowing the other independent variable to account for all of the variance in y that it can account for: that is, the increment in the proportion of variance accounted for in y due to the addition of x 2 to the model: R y ∙ 12 2 − r y 1 2 = . 314 − . 160 = . 154 . This representation is another modification in the traditional Venn diagram approach.

I now show how the modified Venn diagram approach allows us to visualize the basic statistics that are typically reported in the output of multiple regression programs. Two measure of multicollinearity for the i th independent variable are the tolerance and the variance inflation factor: the tolerance equals ( 1 − r 12 2 ) or ( area a ) in the diagram in the first panel and the reciprocal of the tolerance “the variance inflation factor” equals 1 / ( 1 − r 12 2 ) or 1/ area a . The variance in y accounted for by x 1 and x 2 is represented in the left-hand diagram in the second panel as area b ( a r e a b = R y ∙ 12 2 = . 314 ) ; the cross-hatched area in the left-hand diagram. The black area in that diagram is the variance in y not accounted for by the two independent variables and is labelled as area c ( a r e a c = 1 − R y ∙ 12 2 = . 686 ) . These statistics are often reported in the output for multiple regression and correspond directly with the areas in the diagrams. Not surprisingly the diagrams that involve areas of overlap relate to squared terms in multiple regression.

β y 2 ∙ 1 2 , is the squared standardized regression coefficient for y regressed on x 2 controlling for x 1 and is equal ( area d / area a ); that is, the increment in the proportion of the variance in the dependent variable accounted for when x 2 is added to a model that contains the other independent variable divided by the proportion of the variance in x 2 that is independent of the other independent variable. It is the rate of change in the proportion of the variance in y accounted for uniquely by x 2 for a change that is equal to the proportion of the variance in x 2 that is independent of the other independent variable:

For our data, taking the square root of β y 2 ∙ 1 2 yields the absolute value of the standardized regression coefficient: | β y 2 ∙ 1 | = . 154 / . 510 = . 550 . Eq ( 2 ) can be derived from Eq (3.5.7) in [ 1 ].

The F -test for this coefficient (with the null hypothesis that the standardized coefficient is zero) is the proportion of the variance in y that is uniquely accounted for by x 2 divided by the proportion of the variance in y variable that is not accounted for by the independent variables that has itself been divided by its associated degrees of freedom. This provides a significance test for β y 2 ∙ 1 2 , or for β y 2∙1 , or (as we will see) the unstandardized regression coefficient b y 2∙1 :

Where F has one degree of freedom associated with the numerator and ( n − k − 1) degrees of freedom associated with the denominator.

The standard error for the standardized regression coefficient ( β y 2∙1 ) is the square root of the proportion of the variance in y not accounted for by the independent variables in the model that has been divided by its associated degrees of freedom divided by the proportion of the variance in x 2 that is independent of the other independent variable; that is, the square root of area c divided by its associated degrees of freedom divided by the square root of area a :

These components for these inferential statistics are distinctly visualized in the modified Venn diagrams and offer clear intuitions into how these inferential statistics work. For example, for the F -test area d is the unique proportion of the variance in y associated with x 2 (the unique explanatory power of x 2 ) and area c is the proportion of the variance in y unaccounted for by independent variables ( 1 − R y ∙ 1,2 2 ) : the residual or “error” proportion of the variance of y . The larger the unique proportion of variance accounted for by x 2 and the smaller the proportion of the variance unaccounted for by the independent variables, the greater the calculated value of F (all else remaining the same). Importantly, area c in the denominator is divided by its degrees of freedom (to become the residual variance) showing the effects of the sample size and the number of independent variables on the calculated value of F in the Eq ( 4 ). One can easily visualize the effects of making area c smaller and area d larger (and other combinations) on the results of an F -test.

The most common test of the statistical significance of the regression coefficients is to use a t -test that calculates the value of t as the regression coefficient divided by its standard error:

When there is one degree of freedom associated with the numerator t equals F . One will get the same result using t or F to test for the statistical significance of β y 2 ∙ 1 2 , or β y 2∙1 , or the unstandardized regression coefficient b y 2∙1 . Interestingly, area a , which represents the tolerance, drops out of the significance test. It represents multicollinearity, and its absence shows that the value of this measure of multicollinearity and its reciprocal (VIF) do not affect the significance test for the regression coefficients.

Diagrams when there are k independent variables

The extension to the k independent variable situation is straightforward for the modified Venn diagram approach. First, however, a quick note on notation is in order: R y ∙ 1,2 ⋯ k 2 is the multiple correlation coefficient squared for y regressed on all k of the independent variables. R y ∙ 1,2 ⋯ ( i ) ⋯ k 2 is the multiple correlation coefficient squared for y regressed on all k of the independent variables except for the i th independent variable, which can be any one of the k independent variables. R i ∙ 1,2 ⋯ ( i ) ⋯ k 2 is the multiple correlation coefficient squared between the i th independent variable and the other independent variables in model. The ( i ) notation, which indicates the absence of the i th variable in the list, is also used with standardized regression coefficients β y i ∙ 1,2 ⋯ ( i ) ⋯ k 2 . In Fig 4 , we treat x 5 (the fifth independent variable) as the i th independent variable, so that i will represent this fifth independent variable (the independent variable of interest) in that figure.

Table 1 contains the correlation matrix that we work with to construct the overlapping areas in the diagrams in Fig 4 . In this example, I calculate the basic statistics that were presented in the previous section, and I show how the analysis of variance of regression corresponds to the modified Venn diagrams.

The first panel of Fig 4 shows the area of overlap of x 5 with the other four independent variables by regressing x 5 on the other independent variables: R i ∙ 1,2 ⋯ ( i ) ⋯ k 2 . The overlap is .477, which means that the proportion of the variance of x 5 that is linearly independent of the other independent variable is .523. This leads again to two common measures of multicollearity: tolerance: 1 − R i ∙ 1,2 ⋯ ( i ) ⋯ k 2 and the variance inflation factor 1 / ( 1 − R i ∙ 1,2 ⋯ ( i ) ⋯ k 2 ) or in terms of the diagrams ( area a = .523) and (1/ area a = 1.91), respectively. The smaller the independent variance, area a , is; the smaller the tolerance and the greater the variance inflation factor.

The left-hand side of the second panel shows that the proportion of variance in y associated with the independent variables is .365 ( R y ∙ 1,2 ⋯ k 2 ), the cross-hatched area: area b . In this case the proportion of the variance that is not accounted for in y ( 1 − R y ∙ 1,2 ⋯ k 2 ) is .635 (= 1 − .365): area c . The diagram on the right-hand side of this panel shows two components of interest. When we regress the dependent variable on all of the independent variables except for x 5 , we find that the proportion of variance in y accounted for by the four other independent variables is .271: R y ∙ 1,2 ⋯ ( i ) ⋯ k 2 . This means that the proportion of variance in y accounted for by x 5 uniquely is .094 (= .365 − .271). This is labelled area d and is the increment in the proportion of variance in y accounted for by adding x 5 to a model that contains the other independent variables: R y ∙ 1,2 ⋯ k 2 − R y ∙ 1,2 ⋯ ( i ) ⋯ k 2 . Below, I use these four labelled areas from the modified Venn diagrams to show how these areas relate to the most commonly presented results from standardized and unstandardized regression analyses. The squared standardized regression coefficient for the x 5 is:

The absolute value of the standardized regression coefficient (the square root of β y i ∙ 1,2 ⋯ ( i ) ⋯ k 2 ) is .423. We can use the F -test to determine the statistical significance of this result (the null hypothesis is that the coefficient is zero in the population), and for this example I will assume the correlation matrix is based on 106 cases:

The standard error is:

The confidence interval is then constructed by looking up a critical value in a t -table with the appropriate degrees of freedom and alpha level and calculating:

Both the t -test and F -test give the same result in terms of statistical significance when there is one degree of freedom associated with the numerator, in which case t = F . To show this is the case, we divide the standardized regression coefficient by its standard error to produce the calculated value of t and then compare this to the F -test, we use to test for the statistical significance of the regression coefficient:

Typically the output from a multiple regression analysis includes the regression coefficients, their standard errors, their t -values, and confidence intervals. All of these are presented above and their components presented diagrammatically (except for critical values of t and F ).

We can transform the results to unstandardized values by multiplying the standardized regression coefficients by sd y / sd i where sd y is the standard deviation of the dependent variable and sd i is the standard deviation of the i th independent variable:

The standard error for the unstandardized regression coefficients is derived similarly:

The F -test, t -test, and various R 2 values are the same whether we use standardized or unstandardized regression analysis, while the regression coefficients and their standard errors differ depending on whether standardized or unstandardized regression is used.

One other set of statistics that are frequently accompany the output from regression analysis programs is the analysis of variance of regression table ( Table 2 ). The figure corresponding to this table is Fig 4 .

Multiplying the areas by the Σ y 2 provides the unstandardized sums of squares accounted for (model sums of squares) and not accounted for (error sums of squares) in y and the total sums of squares. The mean squares are the sums of squares divided by the corresponding degrees of freedom, and F is the mean square associated with the model divided by the mean square for error. The Σ y 2 cancel each other out in the computation of F in the final column. This F -test has a null hypothesis that the independent variables in the model account for none of the variance in y in the population and has k degrees of freedom associated with the numerator. The F -test used to test for the significance of a single partial regression coefficient ( Eq 7 ) had only one degree of freedom associated with the numerator.

The use of Venn diagrams has been suggested in the literature because of they allow students and researchers to “see” diagrammatically many or the key components in multiple regression [ 1 , 4 , 5 , 10 ]. These diagrams are not seen as a replacement for the algebra (or calculus or matrix algebra) associated with regression analysis, but as an additional tool to help students and researchers gain a better intuitive understanding of these methods. In the multiple regression context; the traditional Venn diagram approach is helpful in some cases, but misleading in others.

Problems arise for the traditional Venn diagram approach used in statistics when there are two or more independent variables. In this situation the traditional Venn diagram approach fails to adequately represent many components as areas of overlap for the variables in the Venn diagrams: and it is these overlaps that are essential to understanding visually the components of multiple regression. Because of these types of problems some suggest doing away with the traditional Venn diagram representations. Hunt [ 11 ] has sections in his article on the design of ballentines entitled “What Ails Ballentines Representing Partial and Multiple Correlations and another entitled, “Depicting Suppressor Variables: A Fatal Ailment.” (The “ballentine” is an alternative expression for the traditional Venn diagram approach used in statistics.) Fox [ 12 ] notes that the overlaps in Venn diagrams would have to be negative to adequately represent some situations.

As an alternative, I propose a modified Venn diagram approach that considers only the overlap between two Venn circles at any one time. It represents the overlap of two or more independent variables with the dependent variable with only a single circle for those independent variables and the overlap of multiple independent variables with another independent variable with only a single circle used to represent the multiple independent variables. The circles are of equal size to represent the fact that a combination of variables can account for a proportion of variance in the target variable that ranges from 0.00 to 1.00.

I am not the first to use a single circle to represent the combined relationships of several variables with another variable, but the insistence on using only two-circle Venn diagrams, the reasoning behind using equal size circles, and the concept of a target variable are new (certainly new in combination with one another). The modified Venn diagram approach allows a consistent diagrammatic representation of shared variance for an independent variable with other independent variables and for the dependent variable with independent variables.

This approach allows students to have a consistent diagrammatic representation of regression coefficients, their standard errors, and F -tests and t -tests that determine their statistical significance. There is more that can be represented, but I leave other extensions to the interested reader. Many of these extensions will be straightforward using the approach taken in this paper.

Funding Statement

The author received no specific funding for this work.

Data Availability

• Subject guides
• Researching for your literature review
• Develop a search strategy

Researching for your literature review: Develop a search strategy

• Literature reviews
• Literature sources
• Before you start
• Keyword search activity
• Subject search activity
• Combined keyword and subject searching
• Online tutorials
• Apply search limits
• Run a search in different databases
• Supplementary searching
• Save your searches
• Manage results

Identify key terms and concepts

Start developing a search strategy by identifying the key words and concepts within your research question. The aim is to identify the words likely to have been used in the published literature on this topic.

For example: What are the key infection control strategies for preventing the transmission of Meticillin-resistant Staphylococcus aureus (MRSA) in aged care homes .

Treat each component as a separate concept so that your topic is organised into separate blocks (concepts).

For each concept block, list the key words derived from your research question, as well as any other relevant terms or synonyms that you have found in your preliminary searches. Also consider singular and plural forms of words, variant spellings, acronyms and relevant index terms (subject headings).  

As part of the process of developing a search strategy, it is recommended that you keep a master list of search terms for each key concept. This will make it easier when it comes to translating your search strategy across multiple database platforms. 

Concept map template for documenting search terms

Combine search terms and concepts

Boolean operators are used to combine the different concepts in your topic to form a search strategy. The main operators used to connect your terms are AND and OR . See an explanation below:

• Link keywords related to a single concept with OR
• Linking with OR broadens a search (increases the number of results) by searching for any of the alternative keywords

Example: nursing home OR aged care home

• Link different concepts with AND
• Linking with AND narrows a search (reduces the number of results) by retrieving only those records that include all of your specified keywords

Example: nursing home AND infection control

• using NOT narrows a search by excluding results that contain certain search terms
• Most searches do not require the use of the NOT operator

Example: aged care homes NOT residential homes will retrieve all the results that include the words aged care homes but don't include the words residential homes . So if an article discussed both concepts this article would not be retrieved as it would be excluded on the basis of the words residential homes .

See the website for venn diagrams demonstrating the function of AND/OR/NOT:

Combine the search terms using Boolean

Advanced search operators - truncation and wildcards

By using a truncation symbol you can capture all of the various endings possible for a particular word. This may increase the number of results and reduce the likelihood of missing something relevant. Some tips about truncation:

• The truncation symbol is generally an asterisk symbol * and is added at the end of a word.
• It may be added to the root of a word that is a word in itself. Example: prevent * will retrieve prevent, prevent ing , prevent ion prevent ative etc. It may also be added to the root of a word that is not a word in itself. Example: strateg * will retrieve strateg y , strateg ies , strateg ic , strateg ize etc.
• If you don't want to retrieve all possible variations, an easy alternative is to utilise the OR operator instead e.g. strategy OR strategies. Always use OR instead of truncation where the root word is too small e.g. ill OR illness instead of ill*

There are also wildcard symbols that function like truncation but are often used in the middle of a word to replace zero, one or more characters.

• Unlike the truncator which is usually an asterisk, wildcards vary across database platforms
• Common wildcards symbols are the question mark ? and hash #.
• Example:  wom # n finds woman or women, p ? ediatric finds pediatric or paediatric.  

See the Database search tips for details of these operators, or check the Help link in any database.

Phrase searching

For words that you want to keep as a phrase, place two or more words in "inverted commas" or "quote marks". This will ensure word order is maintained and that you only retrieve results that have those words appearing together.

Example: “nursing homes”

There are a few databases that don't require the use of quote marks such as Ovid Medline and other databases in the Ovid suite. The Database search tips provides details on phrase searching in key databases, or you can check the Help link in any database.

Subject headings (index terms)

Identify appropriate subject headings (index terms).

Many databases use subject headings to index content. These are selected from a controlled list and describe what the article is about. 

A comprehensive search strategy is often best achieved by using a combination of keywords and subject headings where possible.

In-depth knowledge of subject headings is not required for users to benefit from improved search performance using them in their searches.

Advantages of subject searching:

• Helps locate articles that use synonyms, variant spellings, plurals
• Search terms don’t have to appear in the title or abstract

Note: Subject headings are often unique to a particular database, so you will need to look for appropriate subject headings in each database you intend to use.

Subject headings are not available for every topic, and it is best to only select them if they relate closely to your area of interest.

MeSH (Medical Subject Headings)

The MeSH thesaurus provides standard terminology, imposing uniformity and consistency on the indexing of biomedical literature. In Pubmed/Medline each record is tagged with  MeSH  (Medical Subject Headings).

The MeSH vocabulary includes:

• Represent concepts found in the biomedical literature
• Some headings are commonly considered for every article (eg. Species (including humans), Sex, Age groups (for humans), Historical time periods)
• attached to MeSH headings to describe a specific aspect of a concept
• describe the type of publication being indexed; i.e., what the item is, not what the article is about (eg. Letter, Review, Randomized Controlled Trial)
• Terms in a separate thesaurus, primarily substance terms

Create a 'gold set'

It is useful to build a ‘sample set’ or ‘gold set’ of relevant references before you develop your search strategy..

Sources for a 'gold set' may include:

• key papers recommended by subject experts or supervisors
• citation searching - looking at a reference list to see who has been cited, or using a citation database (eg. Scopus, Web of Science) to see who has cited a known relevant article
• results of preliminary scoping searches.

The papers in your 'gold set' can then be used to help you identify relevant search terms

• Look up your 'gold set' articles in a database that you will use for your literature review. For the articles indexed in the database, look at the records to see what keywords and/or subject headings are listed.

The 'gold set' will also provide a means of testing your search strategy

• When an article in the sample set that is also indexed in the database is not retrieved, your search strategy can be revised in order to include it (see what concepts or keywords can be incorporated into your search strategy so that the article is retrieved).
• If your search strategy is retrieving a lot of irrelevant results, look at the irrelevant records to determine why they are being retrieved. What keywords or subject headings are causing them to appear? Can you change these without losing any relevant articles from your results?
• Information on the process of testing your search strategy using a gold set can be found in the systematic review guide

Example search strategy

A search strategy is the planned and structured organisation of terms used to search a database.

An example of a search strategy incorporating all three concepts, that could be applied to different databases is shown below:

You will use a combination of search operators to construct a search strategy, so it’s important to keep your concepts grouped together correctly. This can be done with parentheses (round brackets), or by searching for each concept separately or on a separate line.

The above search strategy in a nested format (combined into a single line using parentheses) would look like:

("infection control*" OR "infection prevention") AND ("methicillin resistant staphylococcus aureus" OR "meticillin resistant staphylococcus aureus" OR MRSA) AND ( "aged care home*" OR "nursing home*")

• << Previous: Search strategies - Health/Medical topic example
• Next: Keyword search activity >>

• Request new password
• Create a new account

The Essential Guide to Doing Your Research Project

Student resources, chapter 6: reviewing literature.

A.    Checklist for Writing a Literature Review

             Have you:

Read quite a few good, relevant reviews Identified the variables in your study Developed a list of synonyms or alternates Placed variables in a Venn diagram Compiled/located citations with abstracts Read abstracts and culled all irrelevant articles Assessed whether you need to dig deeper or focus your review Systematically logged your relevant readings Read and annotated each relevant article Sorted and organized your annotations Developed a potential outline for your literature review Written purposefully Used the literature to back up your arguments Adopted an appropriate style and tone Gotten plenty of feedback Redrafted (maybe several times)

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

A consistent and general modified Venn diagram approach that provides insights into regression analysis

Roles Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation Department of Sociology, University of Oregon, Eugene, Oregon, United States of America

• Robert M. O’Brien

• Published: May 17, 2018
• https://doi.org/10.1371/journal.pone.0196740
• Reader Comments

Venn diagrams are used to provide an intuitive understanding of multiple regression analysis and these diagrams work well with two variables. The area of overlap of the two variables has a one-to-one relationship to the squared correlation between them. This approach breaks down, however, with three-variables. Making the overlap between the pairs of variables consistent with their squared bivariate correlations often results in the overlap of two of these variables with the third variable that is not the same as the variance of the third variable accounted for by the other two variables. I introduce a modified Venn diagram approach that examines the relationships in multiple regression by using only two circles at a time, provides a new and consistent reason why the circles need to be of the same size, and designates a “target variable” whose overlap with the other circle corresponds to the variance accounted for by the other variable or variables. This approach allows the visualization of the components involved in multiple regression coefficients, their standard errors, and the F -test and t -test associated with these coefficients as well as other statistics commonly reported in the output of multiple regression programs.

Citation: O’Brien RM (2018) A consistent and general modified Venn diagram approach that provides insights into regression analysis. PLoS ONE 13(5): e0196740. https://doi.org/10.1371/journal.pone.0196740

Editor: Fengfeng Zhou, Jilin University, CHINA

Received: January 2, 2018; Accepted: April 19, 2018; Published: May 17, 2018

Copyright: © 2018 Robert M. O’Brien. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper.

Funding: The author received no specific funding for this work.

Competing interests: The author has declared that no competing interests exist.

Introduction

The use of Venn diagrams in statistics can provide a way to make concepts such as variance accounted for in the dependent variable, multiple regression coefficients, the effects of multicollinearity between the independent variables on standard errors, and associated significance tests more intuitive to students and professionals [ 1 – 6 ]. The traditional Venn diagram approach does this by making an analogy between the proportion of area of overlap between circles that represent two variables and the proportion of variance accounted for. This works clearly and simply for bivariate regression, but it only works sometimes in situations with two or more independent variables. This inconsistency creates problems for using the traditional Venn diagram approach to represent multiple regression problems where there are two or more independent variables.

This inconsistency limits the usefulness of the traditional Venn diagram approach. In this paper a modified Venn diagram approach is outlined that addresses the problems with the traditional approach when there are two or more independent variables. I carefully lay out the modified Venn diagram approach and show how these diagrams relate to the most commonly reported statistics associated with multiple regression. I show how the diagrams relate to the formulas for these statistics.

Two examples from the traditional approach

The traditional Venn diagram approach works in the bivariate case: one independent variable and one dependent variable. For convenience of representation, each variable has the same variance and same area. This is accomplished by using standardized variables, so that each variable has a variance of one.

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0196740.g001

This approach works well in the two-variable case (two Venn circles). Note that the diagram represents squared quantities such as squared correlations, the proportion of the variance (a squared quantity) that is accounted for by the independent variable, the proportion of the variance that is not accounted for by the independent variable, and the squared standardized regression coefficient.

https://doi.org/10.1371/journal.pone.0196740.g002

The variance shared by the two independent variables is .49 as implied by the diagram, but this shared variance for the independent variables can be misleading in the case of three independent variables. With three independent variables; for example, suppression can occur among these variables and make the total overlap of the independent variables based on their bivariate relationships to each other misleading.

Two independent and one dependent variable using a modified Venn diagram approach

In Fig 3 , I present a modified Venn diagram with the same data used in Fig 2 , but this diagram is consistent in terms of overlaps and the variance accounted for in the “target variables.” I first show how this modified Venn diagram approach can be used to illustrate the regression components of most interest in the two independent variable case. In the following section, I generalize this approach to k independent variables. Achieving consistency between the proportions of variance accounted for and proportion of area overlapped requires modifications in the traditional approach to using Venn diagrams for statistical interpretations. These modifications include: (1) The number of circles in the diagram considered at any one time is no greater than two (see [ 2 ]) for an excellent example of what can be accomplished by concentrating on just two circles in a Venn diagram]. (2) Each of the circles has the same size, but this is not because they are standardized to have variances of 1.00. They are of the same size to represent that each variable or combinations of variables can account for a proportion of variance in the “target” variable of from 0 to 1.00. (3) In my two circle diagrams, one of the circles represents one or more variables and the other circle is considered the target variable: the target variable is the one for which the proportion of the area overlapped by the other circle is proportional to the variance accounted for in that variable. (4) We assign the overlap of the dependent variable with the combined independent variables into a portion that is uniquely associated with the independent variable of interest (the target independent variable) and a portion that is not uniquely associated with the target variable.

https://doi.org/10.1371/journal.pone.0196740.g003

Circles of the same size allow for the proportion of the area overlapped in the target variable to correspond to the proportion of the variance accounted for. This innovation in the rationale for using circles of the same size is necessary because the circle representing the combined effects of x 1 and x 2 , for example, is not a standardized variable with a variance of one. It is of the same size as the target variable so that it can overlap in area with the target variable from 0 to 1.00, which corresponds to the variance in the target variable that can be statistically associated with one or more variables. The modified Venn diagram approach provides a diagrammatic visualization of the statistical components that provide all of the standard output for basic multiple regression analyses. I focus on the basic regression analysis components to illustrate the modified Venn diagram approach.

Fig 3 contains two panels or rows of diagrams. The top panel displays the overlap of the two independent variables, the target variable (the one for which we compute the regression coefficient and associated statistics: x 2 ). This diagram is easily handled by the traditional Venn diagram approach or this modified approach. The overlap of the target variable with x 1 is .49 in the diagram and .51 of the target variables’ variance is independent of the other independent variable or not accounted for by the other independent variable in the model. We have labelled this independent or unaccounted for variance in the target independent variable as “ area a ” in the diagram.

Diagrams when there are k independent variables

https://doi.org/10.1371/journal.pone.0196740.g004

Table 1 contains the correlation matrix that we work with to construct the overlapping areas in the diagrams in Fig 4 . In this example, I calculate the basic statistics that were presented in the previous section, and I show how the analysis of variance of regression corresponds to the modified Venn diagrams.

https://doi.org/10.1371/journal.pone.0196740.t001

Typically the output from a multiple regression analysis includes the regression coefficients, their standard errors, their t -values, and confidence intervals. All of these are presented above and their components presented diagrammatically (except for critical values of t and F ).

The F -test, t -test, and various R 2 values are the same whether we use standardized or unstandardized regression analysis, while the regression coefficients and their standard errors differ depending on whether standardized or unstandardized regression is used.

One other set of statistics that are frequently accompany the output from regression analysis programs is the analysis of variance of regression table ( Table 2 ). The figure corresponding to this table is Fig 4 .

https://doi.org/10.1371/journal.pone.0196740.t002

Multiplying the areas by the Σ y 2 provides the unstandardized sums of squares accounted for (model sums of squares) and not accounted for (error sums of squares) in y and the total sums of squares. The mean squares are the sums of squares divided by the corresponding degrees of freedom, and F is the mean square associated with the model divided by the mean square for error. The Σ y 2 cancel each other out in the computation of F in the final column. This F -test has a null hypothesis that the independent variables in the model account for none of the variance in y in the population and has k degrees of freedom associated with the numerator. The F -test used to test for the significance of a single partial regression coefficient ( Eq 7 ) had only one degree of freedom associated with the numerator.

The use of Venn diagrams has been suggested in the literature because of they allow students and researchers to “see” diagrammatically many or the key components in multiple regression [ 1 , 4 , 5 , 10 ]. These diagrams are not seen as a replacement for the algebra (or calculus or matrix algebra) associated with regression analysis, but as an additional tool to help students and researchers gain a better intuitive understanding of these methods. In the multiple regression context; the traditional Venn diagram approach is helpful in some cases, but misleading in others.

Problems arise for the traditional Venn diagram approach used in statistics when there are two or more independent variables. In this situation the traditional Venn diagram approach fails to adequately represent many components as areas of overlap for the variables in the Venn diagrams: and it is these overlaps that are essential to understanding visually the components of multiple regression. Because of these types of problems some suggest doing away with the traditional Venn diagram representations. Hunt [ 11 ] has sections in his article on the design of ballentines entitled “What Ails Ballentines Representing Partial and Multiple Correlations and another entitled, “Depicting Suppressor Variables: A Fatal Ailment.” (The “ballentine” is an alternative expression for the traditional Venn diagram approach used in statistics.) Fox [ 12 ] notes that the overlaps in Venn diagrams would have to be negative to adequately represent some situations.

As an alternative, I propose a modified Venn diagram approach that considers only the overlap between two Venn circles at any one time. It represents the overlap of two or more independent variables with the dependent variable with only a single circle for those independent variables and the overlap of multiple independent variables with another independent variable with only a single circle used to represent the multiple independent variables. The circles are of equal size to represent the fact that a combination of variables can account for a proportion of variance in the target variable that ranges from 0.00 to 1.00.

I am not the first to use a single circle to represent the combined relationships of several variables with another variable, but the insistence on using only two-circle Venn diagrams, the reasoning behind using equal size circles, and the concept of a target variable are new (certainly new in combination with one another). The modified Venn diagram approach allows a consistent diagrammatic representation of shared variance for an independent variable with other independent variables and for the dependent variable with independent variables.

This approach allows students to have a consistent diagrammatic representation of regression coefficients, their standard errors, and F -tests and t -tests that determine their statistical significance. There is more that can be represented, but I leave other extensions to the interested reader. Many of these extensions will be straightforward using the approach taken in this paper.

• 1. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. Hillside NJ: Lawrence Erlbaum Associates; 1983.
• View Article
• PubMed/NCBI
• Google Scholar
• 12. Fox J. Applied regression analysis and generalized linear models, 3rd ed. Thousand Oaks, CA: Sage; 2008.

• Publishing Policies
• For Organizers/Editors
• For Authors
• For Peer Reviewers

Using The Venn Diagram For Developing University Students’ Analytical Geographical Thinking

Developing university students’ analytical geographical thinking is one of the major objectives of university geographical education irrespective of the fact that it is achieved during Didactics of Geography lectures and seminars or fundamental courses in Geography. This objective can be achieved by various means, the Venn Diagram being among them. It is used to identify the aspects characteristic to concepts or the common features of geographical systems. Our research raised the following questions: which are the main geographic concepts and systems as the object of a certain comparison, and second, what are the main criteria on which a comparison is completed? To answer these questions, we used Venn Diagrams as research material made by students and published in articles, chapters, books, or as part of their portfolios for learning Geography and the Didactics of Geography, as well as during written exams or as exercises included in lesson planning. Our conclusions are the following: the critical and analytical geographical thinking have a peripheral position within the educational process, many sampled diagrams highlight the comparison of different objects without specifying the particular criteria, and students make use of critical and analytical thinking but with certain limitations when the geographic language is used. Keywords: Competence Didactics of Geography comparison criteria higher education graphical organizer cognition

Introduction

One of the major objectives of university geographical education is developing university students’ analytical geographical thinking. This is true either for Didactics of Geography lectures and seminars or for fundamental courses in geography. There is well acknowledged that in territorial planning research three stages have to be completed in order to investigate all spatial features through the lens of systemic analysis: the human and economic analysis and the global investigation of all features occurred from the previous approach ( Zotic, 2005, pp. 31-32 ).

The geographic analysis represents a mental or a real action through which the whole territorial system is decomposed into its elements, each being examined to identify its properties and to establish the relationship between them. In this regard, each element is approached to find out the main relations established between all the components of a territorial system. Studying each part of a system is important to exclusively separate all the parts, and the investigation upon them to be undoubtedly complete. Accordingly, this approach is meant to unveil both the internal structure of the investigated spatial elements as well as their specificities ( Dulamă, 2010a, p. 163 ). The second step of the critical thinking is the comparison, thus enabling the mental or physical approach to geographical objects, systems, and processes in order to identify the main similarities and differences among them. Against such a background, the comparison is completed based on particular criteria and, in line with each criterion, there are multiple opportunities to discover whether the compared objects are similar or different ( Dulamă, 2010a, p. 163 ).

In geographical research, there are various means for student’ critical and analytical thinking development, through Venn Diagram usage. It was created by the English logician John Venn (1834-1923) to represent visually the complex logical propositions and algebraic statements ( Edwards, 2004 ). Canela Morales and Ruiz Sosa ( 2020 ) pointed out in a thematic study the symbolic nature of formal rationale of this diagram and its diagrammatic nature, as well as how knowledge could be ensured using diagrams and symbols. The authors investigated the various ways of diagrams use in the teaching and learning processes.

Venn and Euler Diagrams are well-defined mathematical diagram types used during Mathematics examinations in secondary education in the UK. In order to make the grading more efficient, an automatic system for assessing student’s answers was created based on Venn and Euler Diagrams. The grade is assigned after comparing the student’s answer with a model answer ( Wijesinghe et al., 2017 ).

Focused on the use of Venn Diagram in Geography, the whole body of the international literature in the field introduces only indirectly the contribution of this didactic resource during Geography teaching and learning. Some aspects highlight the wide range of the various forms of graphical organizers ( see Gottfried, 2015 ), while other studies are concerned with the natural sciences ( Lynam et al., 2007 ). In Romania, Venn Diagram started to be extensively used in geography education after 2000, as a result of the awareness of its importance during training programs in which geography teachers were involved. Although this diagram is presented in some logic studies, several teachers have learned in recent decades that the Venn Diagram is a cognitive organizer of two partially overlapping circles that represent the similarities and differences between two aspects, ideas, or concepts ( Steele et al., 1998a ; 1998b ). The similarities are mentioned in the area where the circles overlap, and the differences in the free, outer areas ( Dulamă, 2002, p. 166 ). When using the Venn Diagram, spatial information is classified into two categories: real objects and their spatial properties and abstract objects ( Gottfried, 2015 ).

Problem Statement

Although in Romania the Venn Diagram is largely introduced by a whole body of the literature focused on the Didactics of Geography, its benefits for the critical thinking development remain peripheral. Furthermore, this cognitive organizer is underused in the Romanian educational system both in pre-university education and in academic training. A review of the observed educational practices, pointing out a certain teaching style ( Fetti & Albulescu, 2020 ), and of the specialized literature shows that there are different approaches on Venn Diagrams, thus generating relevant qualitative differences in line with its efficiency in the students’ critical thinking, who are interested in learning Geography. Both the literature and various practices show that some teachers have no relevant information on the role of the diagram within knowledge modelling in a certain domain ( Gottfried, 2015 ).

Research Questions

Using the Romanian literature review concerned with the Venn Diagram, we intend to find some answers to the following questions: which are the main geographic concepts and systems as objects of a certain comparison, and second, what are the main criteria on which a comparison is completed? In addition, the study aims to reveal the main compared aspects, attributes and objects, the contexts and situations in which students use comparisons according to certain criteria. Finally, based on these findings, we are interested in unveiling some information on how students use Venn Diagrams in their particular learning actions, as well as on students’ difficulties in certain situations of comparing geographic objects, ideas, elements, and geographic processes.

Purpose of the Study

This research focuses on the following major objectives: on the one hand, it aims to investigate different Venn Diagrams included in the Romanian specialized literature to understand their main theoretical and methodological perspectives and, on the other hand, to analyze Venn Diagrams made by the students to find out the ways they use to include Venn Diagrams in their own learning processes, as well as to identify the students’ major difficulties within these specific training situations.

Research Methods

Procedure . Within the electronic portfolio made for the subject entitled “Geography, environment/sciences and their didactics in kindergarten and the primary grades”, during the Spring semester of the 2020, the students had to compare two distinct major landforms on various continents (e.g. mountains, hills, plateaus and fields), by completing two particular aspects of a landform, two specificities of the other landform and two common aspects for both investigated landforms within the Venn Diagram. As a background in solving this task, the students were informed about the theoretical support in terms of a handbook, a thematic presentation, four profiles of the investigated landforms and a table including the key aspects of the relief (Table 1 ). The portfolio tasks had to be solved individually until the exam. The work was designed in groups, to avoid any stress caused by the Romanian pandemic crisis.

Participants . 37 students in their 3 rd year of study, attending the Program of the Primary Education Pedagogy from the Faculty of Psychology and Education Sciences, Babeş-Bolyai University from Cluj-Napoca, Romania, were involved in the research. These students train for a teaching career in the primary educational field of the Romanian pre-university education system. Of these, about 50% graduated the Pedagogic College and work in the primary education system. The second author of this study assigned this task to the students, assessed all students’ portfolios, being perceived by the students as a professor, not as a researcher.

The research material comprised 20 Venn Diagrams included in various works on Didactics of Geography ( Dulamă, 2002 ; 2008 ; 2009 ; 2010a ; 2010b ; 2010c ; Ilovan et al., 2010 ; Ilovan & Mihalca, 2010 ; Ţolaş, 2010 ) and in two scientific papers ( Costa & Antonie, 2006 ; Dulamă & Ilovan, 2004 ), identified by the first author. The research material also covered eight diagrams made by some groups of four students, all diagrams being included in the students’ portfolios in order to be evaluated. The collected data are represented in tables specific to Education Sciences ( Magdaș, 2018 ).

Collecting data and data analysis . Diagrams were extracted both from the specialized literature and from the students’ portfolios. Their content was then critically investigated in terms of their topics and contents.

Literature review on Romanian Venn Diagrams specialized work

Based on the critical analysis of the content included in the previously mentioned 20 diagrams, several aspects have been identified as follows:

a. Concepts and geographic systems are subjected to comparison. Literature recommends that Venn Diagrams are used for diverse comparisons: concepts (river/stream, glacier/ice shelf, island/peninsula, delta/estuary), hydrographic elements (The Rhine/The Danube, The Black Sea/The Caspian Sea, Aral Lake/Baikal Lake), mountains (The Alps/The Carpathians, The Western Carpathians/The Southern Carpathians), countries (Romania/Hungary, Italy/Norway, Romania/France, Italy/Greece, Brazil/Argentina), people (Romanians/Dutch, Hungarians/Romanians, Greeks/Italians, etc.), animal species (cat/lion), and types of plants ( Dulamă, 2008, p. 346 ). In other works, diagrams are used to compare continents, countries, cities, lakes, civilizations, and types of forests ( Dulamă & Mihalca, 2010 ).

b. Key features using Venn Diagrams. A sample of each identified category has been extracted (Table 2 ) to analyze through the lens of the critical geographic thinking. Differences occur after comparing all these diagrams. Some of them have been made by various authors to provide diverse examples for teachers and students (Dulamă, 2002, 2008, 2009), while others were achieved by the students in different contexts for their work assessment ( Dulamă & Ilovan, 2004 ) or in various research contexts ( Costa & Antonie, 2006 ). A diagram includes information about the people and works as a subjective tool/resource ( Dulamă & Ilovan, 2004 ), while the other ones unveil some geographic information with higher objectivity ( Costa & Antonie, 2006 ; Dulamă, 2002 ; 2009 ). They do not include criteria on which the comparison has been made with these issues being implied. Multiple aspects were presented in a mirroring manner; particularly a certain aspect has been analyzed in both compared geographical systems and concepts. Considering diagrams about geographical systems, there have been included geographical names of some realia (The Danube, Bucharest, The Carpathian Mountains, and The Black Sea), as well as some general or spatial features as the relief proportionality and variety, or types of countries, etc.

c. Comparative aspects of Venn Diagrams . To set the main categories of aspects revealed by the use Venn Diagrams included in Table 2 , as well as of other diagrams discussed in this research, the data included in Table 3 was systematized. The table is an example of a working tool that could be further developed according to various and countless topics. The main aspects resulted from the table investigation are as follows. When two geographical objects are compared, as for instance two rivers, the analysis can be done at the level of concepts or of some real rivers (the Rhine and the Danube). Considering the concepts, the analysis is made using some abstract objects with real features that define these concepts ( Dulamă, 2010a, p. 163 ). In the case of real rivers, multiple properties in terms of their characteristics, features, particularities, etc. are focused on by Venn Diagrams, all these being perceived as relevant features defining that real or abstract object. Furthermore, they relate to some features that differentiate one object from another ( Academia Română, 2009 ). When it is about the real rivers, their properties are included despite the concepts or features that are not included in this category.

Table 3 shows some information related to Venn Diagrams during the process of comparing the real objects and the real systems of objects. Based on the investigated diagrams, several types of elements have been designed and each category details several elements discovered during the comparison processes using Venn Diagrams.

Students’ analysis of Venn Diagrams

Considering students’ work and involvement in the context of the distance learning through the Microsoft Teams platform, the students had less support from the tutor. The main activities were represented by their individual study using various support teaching resources during their instruction, alongside other resources as syllabus, PowerPoint presentations, resources uploaded in Microsoft Teams as well as in private discussion groups via Facebook. The students discussed and kept in touch with their professor by online forums. Through all these means and methods, specific to the online environment, we did our best to provide a suitable instruction for students.

The proposed tasks aimed to allow the proper understanding and an objective learning assessment with all these important in the consolidation of a real perception towards diagram contents. Since each students’ team chose for comparison the landforms, mountains and plateaus were compared with the hills and fields (plains). The students had to follow the criteria included in Table 4 . The information provided by the table was added only in three diagrams (30%), showing that only some students understood the task. A diagram unveils the usage of term ‘height’, and the concept of ‘reduced height’ instead of ‘altitude’, as a specific concept in physical geography, is used in another diagram. In the second diagram, the terms ‘lower altitude’ and ‘higher altitude’ are used. Within these diagrams, we identified different types of wording that are not in line with the particular criteria since they do not refer to the features of the landforms, on the contrary, they refer to their current or possible use during the process of geography learning.

The ways in which students compared the landforms showed that they lacked relevant knowledge for teaching geography and they frequently used common or colloquial language rather than a geographical one. However, the students demonstrated the use of their critical thinking, making comparisons based on the suggested criteria.

The study focused on Venn Diagrams used in the field of Geography, for various comparisons, and revealed some relevant concluding remarks. The specialized literature revealed that the main actors in the students’ education are less interested in Venn Diagrams, with this teaching resource remaining marginal in teaching and learning Geography. Accordingly, the critical and analytical geographical thinking also have a peripheral position within the educational process, as this topic remains marginal in the specialized literature. The diagrams included in various works illustrate a major diversity of geographic objects (geographic systems) that are differently compared, using various approaches. Many sampled diagrams highlight the comparison of different objects without specifying particular criteria.

The students’ analysis of diagrams indicated relevant difficulties when the diagrams were used. This is argued by the fact that only 30 percent of them are correctly using the resources recommended by the professor. The textual expressions included in diagrams clearly indicate that the students make use of critical and analytical thinking but with certain limitations when the geographic language is used. These issues could be solved through some major time resources in order to properly learn the geographic contents, either during university study programs or through individual study. Making use of some specific exercises aiming at diagrams construction closely supervised by the professor could represent one action in the learning of geography using Venn Diagrams.

Acknowledgments

The research for this article was supported by a STAR-UBB Institute fellowship (The Institute of Advanced Studies in Science and Technology, belonging to Babeș-Bolyai University of Cluj-Napoca, Romania), won by Professor Maria Eliza Dulamă, Ph.D., during the 2019-2020 academic year (for the April-May 2020 period) and titled Valorificarea unor tehnologii avansate pentru realizarea unor filme didactice destinate predării-învățării în învățământul universitar [Valorising Certain Advanced Technologies to Realise Didactic Films for Teaching-Learning in the University System]. The fellowship was funded through the project 33PFE/2018 (Strategic infrastructure at Babeș-Bolyai University in the context of developing new and smart technology – 2018-2020), which was won through a competition organised in 2018 by the Ministry of Research and Innovation, of Romania.

• Academia Română. (2009). Dicționarul explicativ al limbii române [Explanatory Dictionary of Romanian Language]. Univers Enciclopedic Gold.
• Canela Morales, L. A., & Ruiz Sosa, F. G. (2020). General aspects of symbolic and diagrammatic knowledge: the case of Venn diagrams. ANDAMIOS, 16(41), 63-85.
• Costa, A. F., & Antonie, M. (2006). Organizarea informaţiei prin diagrama Venn [Organising information using Venn diagram]. In M. E. Dulamă, O.-R. Ilovan, & F. Bucilă (Eds.), Tendinţe actuale în predarea şi învăţarea geografiei/Contemporary Trends in Teaching and Learning Geography (vol. 2, pp. 360-364). Clusium.
• Dulamă, M. E. (2002). Modele, strategii şi tehnici didactice activizante cu aplicaţii în geografie [Models, strategies and active didactic techniques applied in geography]. Editura Clusium.
• Dulamă, M. E. (2008). Metodologii didactice activizante [Active didactic methodologies]. Clusium.
• Dulamă, M. E. (2009). Utilizarea organizatorilor grafici în învăţarea geografiei [Using graphic organizers in teaching and learning geography]. Materialele Simpozionului Jubiliar International “Mediul si dezvoltarea durabila”, 70 ani de la fondarea Facultăţii de Geografie, Universitatea de Stat din Tiraspol (pp. 311-316). Editura Labirint.
• Dulamă, M. E. (2010a). Fundamente despre competențe [Basics about Competences]. Presa Universitară Clujeană.
• Dulamă, M. E. (2010b). Republica Moldova [The Republic of Moldavia]. In M. E. Dulamă, & A. I. Mihalca (Eds.), Europa. Proiecte de lecţii pentru clasa a VI-a (pp. 98-101). Presa Universitară Clujeană.
• Dulamă, M. E. (2010c). Republica Ungară [The Republic of Hungary]. In M. E. Dulamă, & A. I. Mihalca (eds.) Europa. Proiecte de lecţii pentru clasa a VI-a (pp. 94-97). Presa Universitară Clujeană.
• Dulamă, M. E., & Ilovan, O.-R. (2004). The evalution of the students at the course of intercultural education through geography. Science Education International, 15(3), 225-239.
• Dulamă, M. E., & Mihalca, A. I. (Eds.) (2010). Europa. Proiecte de lecţii pentru clasa a VI-a [Europe. Lesson Projects for the 6th Grade]. Presa Universitară Clujeană.
• Edwards, A. W. F. (2004). Cogwheels of the mind: the story of Venn diagrams. JHU Press.
• Fetti (Mora), D. F., & Albulescu, I. (2020). Developing emotional resilience in the primary education through the teaching style. In V. Chiș (Ed.), 7th ERD, European Proceedings of Social and Behavioural Sciences, 85, 669-676.
• Gottfried, B. (2015). A comparative study of linear and region based diagrams. Journal of Spatial Information Science, 10, 3-20.
• Ilovan, O.-R., & Mihalca, I. A. (2010). Relieful Europei [The relief of Europe]. In M. E. Dulamă, & A.I. Mihalca (Eds.), Europa. Proiecte de lecţii pentru clasa a VI-a (pp. 16-21). Presa Universitară Clujeană.
• Ilovan, O.-R., Dulamă, M. E., & Mihalca, I. A. (2010). Râurile, fluviile, lacurile. Analiza unui fluviu: Dunărea [Streams, rivers and lakes. The study of the Danube river]. In M. E. Dulamă, & A. I. Mihalca (Eds.), Europa. Proiecte de lecţii pentru clasa a VI-a (pp. 28-37). Presa Universitară Clujeană.
• Lynam, T., De Jong, W., Sheil, D., Kusumanto, T., & Evans, K. (2007). A review of tools for incorporating community knowledge, preferences, and values into decision making in natural resources management. Ecology and Society, 12(1), 5.
• Magdaș, I. (2018). Prezentarea și prelucrarea datelor cercetării în științele educației. Ghid pentru studenți [Presenting and processing research data in the Sciences of Education. A students’ guide]. Presa Universitară Clujeană. http://www.editura.ubbcluj.ro/bd/ebooks/pdf/2276.pdf
• Steele, J. L., Meredith, K. S., & Temple, C. (1998a). Lectura şi scrierea pentru dezvoltarea gândirii critice [Reading and writing in the development of critical thinking], III. Edited by Centrul Educaţia 2000+ within the project “Lectura şi scrierea pentru dezvoltarea gândirii critice”. Bucureşti, Cluj-Napoca.
• Steele, J. L., Meredith, K. S., & Temple, C. (1998b). Promovarea gândirii critice [Promoting critical thinking], II. Edited by Fundaţia Soros şi de Institutul pentru o Societate Deschisă în cadrul proiectului “Lectura şi scrierea pentru dezvoltarea gândirii critice”. Cluj-Napoca.
• Ţolaş, A. (2010). Grecia [Greece]. In M. E. Dulamă, & A. I. Mihalca (Eds.), Europa. Proiecte de lecţii pentru clasa a VI-a (pp. 131-141). Presa Universitară Clujeană.
• Wijesinghe, D. B., Kadupitiya, J., Ranathunga, S., & Dias, G. (2017). Automatic assessment of student answers consisting of Venn and Euler diagrams. 2017 IEEE 17th International Conference on Advanced Learning Technologies, 243-247.
• Zotic, V. (2005). Componentele operaţionale ale organizării spaţiului geografic [Operational Components for Organising the Geographical Space]. Presa Universitară Clujeană.

Copyright information

About this article

Publication date.

31 March 2021

Article Doi

https://doi.org/10.15405/epsbs.2021.03.02.26

978-1-80296-103-4

European Publisher

Print ISBN (optional)

Edition number.

1st Edition

Education, teacher, digital education, teacher education, childhood education, COVID-19, pandemic

Cite this article as:

Cîineanu, M., Dulamă, M. E., Ilovan, O., Rus, G. M., Kobulniczky, B., Voicu, C., & Chiș, O. (2021). Using The Venn Diagram For Developing University Students’ Analytical Geographical Thinking. In I. Albulescu, & N. Stan (Eds.), Education, Reflection, Development – ERD 2020, vol 104. European Proceedings of Social and Behavioural Sciences (pp. 238-247). European Publisher. https://doi.org/10.15405/epsbs.2021.03.02.26

We care about your privacy

We use cookies or similar technologies to access personal data, including page visits and your IP address. We use this information about you, your devices and your online interactions with us to provide, analyse and improve our services. This may include personalising content or advertising for you. You can find out more in our privacy policy and cookie policy and manage the choices available to you at any time by going to ‘Privacy settings’ at the bottom of any page.

Manage My Preferences

You have control over your personal data. For more detailed information about your personal data, please see our Privacy Policy and Cookie Policy .

These cookies are essential in order to enable you to move around the site and use its features, such as accessing secure areas of the site. Without these cookies, services you have asked for cannot be provided.

Third-party advertising and social media cookies are used to (1) deliver advertisements more relevant to you and your interests; (2) limit the number of times you see an advertisement; (3) help measure the effectiveness of the advertising campaign; and (4) understand people’s behavior after they view an advertisement. They remember that you have visited a site and quite often they will be linked to site functionality provided by the other organization. This may impact the content and messages you see on other websites you visit.

• Activity Diagram (UML)
• Amazon Web Services
• Android Mockups
• Block Diagram
• Business Process Management
• Chemical Chart
• Cisco Network Diagram
• Class Diagram (UML)
• Collaboration Diagram (UML)
• Compare & Contrast Diagram
• Component Diagram (UML)
• Concept Diagram
• Cycle Diagram
• Data Flow Diagram
• Data Flow Diagrams (YC)
• Database Diagram
• Deployment Diagram (UML)
• Entity Relationship Diagram
• Family Tree
• Fishbone / Ishikawa Diagram
• Gantt Chart
• Infographics
• iOS Mockups
• Network Diagram
• Object Diagram (UML)
• Object Process Model
• Organizational Chart
• Sequence Diagram (UML)
• Spider Diagram
• State Chart Diagram (UML)
• Story Board
• SWOT Diagram
• TQM - Total Quality Management
• Use Case Diagram (UML)
• Value Stream Mapping
• Venn Diagram
• Web Mockups
• Work Breakdown Structure

Literature Review

Copy of Literature Review

You can easily edit this template using Creately. You can export it in multiple formats like JPEG, PNG and SVG and easily add it to Word documents, Powerpoint (PPT) presentations, Excel or any other documents. You can export it as a PDF for high-quality printouts.

• Flowchart Templates
• Org Chart Templates
• Concept Map Templates
• Mind Mapping Templates
• WBS Templates
• Family Tree Templates
• Network Diagram Templates
• SWOT Analysis Templates
• Genogram Templates
• Activity Diagram
• Class Diagram
• Collaboration Diagram
• Component Diagram
• Data Flow Diagrams(YC)
• Deployment Diagram
• Object Diagram
• Sequence Diagram
• State Chart Diagram
• Use Case Diagram

Related Templates