9 posts categorized "Statistics"

December 05, 2013

How to Cite a Data Set in APA Style

Timothy McAdooby Timothy McAdoo

Whether you’re a "numbers person" or not, if you’re a psychology student or an early-career psychologist, you may find yourself doing some data mining.  Psychologists are increasingly encouraged to provide their data online for other researchers to use and analyze.  And big-data psychologist is one of the hot new jobs in the industry.

Because big data is a big deal, you’ll want to know how to cite a data set.

Reference Example

U.S. Department of Health and Human Services, Substance Abuse and Mental
    Health Services Administration, Office of Applied Studies. (2013). Treatment
    episode data set -- discharges (TEDS-D) -- concatenated, 2006 to 2009
[Data
    set]. doi:10.3886/ICPSR30122.v2

Because this data set has a DOI, the reference includes that DOI. For data sets without a DOI, the URL should be included in the reference, like this: "Retrieved from http://www.icpsr.umich.edu/SDA/SAMHDA/30122-0001/CODEBOOK/conc.htm"

Also note that the name of the data set is italicized. And, a description of the material is included in brackets after the title, but before the ending period, for maximum clarity.

big dataIn-Text Citation Example

The in-text citation for this reference would be "U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Office of Applied Studies (2013)" or "(U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Office of Applied Studies, 2013)."  Of course, if you cite that a number of times, you’ll probably want to abbreviate the author name.

Related Documents

Data sets sometimes have many other documents associated with them (e.g., reports, papers, and analyses about the data; tests and measures used to procure the data; user manuals; code books). When you are citing one of these related items, whether instead of or in addition to the data, be sure to describe the format in brackets after the title. For example, in this example from the APA Style Guide to Electronic References, Sixth Edition, "Data file and code book" is used to describe the format:

Pew Hispanic Center. (2004). Changing channels and crisscrossing cultures:
    A survey of Latinos on the news media
[Data file and code book]. Retrieved
    from http://pewhispanic.org/datasets/

The in-text citation would be "Pew Hispanic Center (2004)" or "(Pew Hispanic Center, 2004)."

For more about big data, you may be interested in these pages on the APA website:

Note: I modified this post on 12/23/2013 to include the DOI of the data set in the first reference example.

November 07, 2013

Is It Sometimes Okay to Begin a Sentence With a Lowercase Letter?

Tyler

 

 

by Tyler Krupa

A basic grammar rule is that the first word in a complete sentence should be capitalized. But do you know how to proceed when a name that begins with a lowercase letter begins a sentence? Or whether it is okay to begin a sentence with a lowercase statistical term (e.g., t test or p value)?

Although the two examples listed above seem to be exceptions to the rule that the first word in a sentence should be capitalized, this is not the case. Note that per APA Style, the first word in a complete sentence should always be capitalized.

So what should you do when you come across the above examples in your writing? Getting it right is simple as long as you remember the following two guidelines (see sections 4.14 and 4.30 in the sixth edition of the Publication Manual):

1. If a name that begins with a lowercase letter begins a sentence, then it should be capitalized.

2. Do not begin a sentence with a lowercase statistical term (e.g., t test or p value), a lowercase abbreviation (e.g., lb), or a symbol that stands alone (e.g., α).

To help illustrate the first guideline, let’s look at the following example:

Correct:

Van Morrison and Smith (2012) interviewed 100 participants . . .

Incorrect:

van Morrison and Smith (2012) interviewed 100 participants . . .

In the example above, even though the usual presentation of the surname van Morrison begins with a lowercase v, it is correct to capitalize the first letter of the surname when the name begins a sentence. However, note that if the surname van Morrison is used later in the sentence or in references/citations, then the lowercase v is retained (e.g., At the conclusion of the participant interviews, van Morrison and Smith . . .). For more information on how to correctly capitalize author names, see the following post to our blog.

Now let’s look at an example that illustrates the second guideline:

Correct:

We used t tests to determine . . .

Incorrect:

t tests were used to determine . . .

Incorrect:

t Tests were used to determine . . .

Incorrect:

T tests were used to determine . . .

Note that in the example above, it is not okay to capitalize the statistical term at the beginning of the sentence because doing so changes the meaning of the statistic. Therefore, in instances such as these, it is necessary to recast the sentence. However, note that it is okay to begin a sentence with a capitalized statistical term (e.g., F tests indicated that . . .). For more information on how to format statistics in your paper, see the following post to our blog.

We hope these examples clear up this point of possible uncertainty. Still have questions? Leave us a comment.

August 25, 2011

The Grammar of Mathematics: Writing About Variables

Chelsea blog 2 by Chelsea Lee

In the social sciences, the worlds of grammar and mathematics intersect, as authors must not only run statistical tests but also write about their results in a clear, consistent way. To help achieve that end, this post focuses on some of the grammar of mathematics: how to introduce and use statistical terms in text when you are reporting your results. 

The sixth edition Publication Manual provides a listing of many mathematical variables and terms that commonly appear in APA Style papers (see Table 4.5 on pp. 119–123). The table below excerpts some of the most common statistics, showing their written-out and abbreviated forms in both the singular and the plural. Following that, we discuss the ins and outs of using them in an APA Style paper. 

 

Written-out form

Abbreviation/symbol

Singular

Plural

Singular

Plural

Cohen’s d

Cohen’s ds

d

ds

degree of freedom

degrees of freedom

df

dfs

F statistic or F value

F statistics or F values

F

Fs 

mean

means

M

Ms

sample size (subsample)

sample sizes (subsample)

n

ns 

sample size (full sample)

sample sizes (full sample)

N

Ns 

p value

p values

p

ps 

r value

r values

r

rs 

R2 value

R2 values

R2 

R2s 

standard deviation

standard deviations

SD

SDs 

standard error

standard errors

SE

SEs 

t value

t values

t

ts 

z score

z scores

z

zs 

Cronbach’s alpha

Cronbach’s alphas

Cronbach’s α

Cronbach’s αs

beta

betas

β

βs

chi-square

chi-squares

χ2

χ2s

delta

deltas

Δ

Δs

Singular Versus Plural

  • The syntax of your sentence will dictate whether you need to use the singular or plural form of the variable.
  • All plural abbreviated forms are made by adding a nonitalic lowercase “s.” Do not use an apostrophe plus an “s,” an italic “s,” or a capital “S.”
    • Correct: ps < .05; Ms = 3.70 and 4.22; degrees of freedom.
    • Incorrect: ps < .05, p’s < .05; Ms = 3.70 and 4.22; Means = 3.70 and 4.22; degree’s of freedom.

Written-Out Form Versus Abbreviated Form

  • Use the written-out form of the variable in prose; use the symbol in conjunction with all mathematical operators (such as the equals sign or the greater than/less than signs).
  • As usual, use singular or plural as needed by the context.

Italic Versus Nonitalic

  • Variables are italicized.
  • Superscript numbers are not italicized (e.g., R2).
  • Identifiers (which can be superscript or subscript words, letters, or numbers) are not italicized. For example, if Mgirls = 4.22 and Mboys = 3.78, the symbol for mean is italicized, but the nonvariable identifiers (here identifying the two groups, “girls” and “boys”) are not italicized.  

An Example

The means and standard deviations are reported in Table 1. We calculated Cronbach’s alpha as the reliability statistic and then ran a chi-square test. The read-aloud group (M = 4.55, SD = 0.65) and the read-silently group (M = 2.72, SD = 0.53) differed significantly on the test of reading comprehension, χ2(1, 50) = 4.25, p < .05. Boys and girls did not differ significantly (Mgirls = 4.22 and Mboys = 3.78). The sample size for each testing group was 25, but several participants in each group (ns = 5 and 6, respectively) had missing data on the final question, and these were replaced with the participant’s mean score. This did not affect reliability (Cronbach’s α = .83). 

December 16, 2010

Journal Article Reporting Standards: How Do They Work?

HCooper 3-1-09

by Harris Cooper, PhD

Last week I discussed why APA’s Journal Article Reporting Standards (the JARS) are needed when you are writing your psychology research report. I compared a psychology research paper to assembly instructions, like those you would follow when constructing a shelf or putting a bike together. Without a list of materials and clear instructions, others will find it difficult to understand what you did and to repeat your experiment. In this post, I describe how the JARS works. (For more detail about why the JARS is needed and how it works, see my book, Reporting Research in Psychology: How to Meet Journal Article Reporting Standards, which was released by APA this week.)

The JARS first focuses on information recommended for inclusion in all reports. These recommendations are organized by the parts of a research report: title, abstract, introduction, Method, Results, and Discussion. For example, here are the characteristics of sampling procedures that the JARS recommends be described in the Method section of every report:

  • Sampling method, if a systematic sampling plan was implemented
  • Percentage of sample approached that participated
  • Self-selection (by either individuals or units, such as schools or clinics)
  • Settings and locations where data were collected
  • Agreements and payments made to participants
  • Institutional review board agreements, ethical standards met, and safety monitoring

If a study is conducted with a college or university subject pool, the description of the sampling procedure is pretty straightforward. But, if a study involves, say, a classroom intervention, you can see how the description can become more involved. In both cases, however, if we don’t know this information about the study, it would be hard to determine to whom the results of the study apply.
After the recommendations that pertain to all reports, the JARS asks that researchers pay careful attention to reporting the research design. Research designs come in a variety of forms depending on the type of question that motivates the research. Therefore, each design requires unique information.
The JARS currently provides standards for only one family of designs, those involving purposive or experimental manipulations or interventions. Among the information requested are these aspects of the manipulation:

  • Setting (where the manipulations or interventions occurred)
  • Exposure quantity and duration (how many sessions, episodes, or events were intended to be delivered and how long they were intended to last)
  • Time span (how long it took to deliver the intervention or manipulation to each unit)
  • Activities to increase compliance or adherence (e.g., incentives)

So, if a study evaluates the effect of providing tutoring for children having difficulty learning to read, the JARS recommends that the report tell readers where the tutoring took place (e.g., in the regular classroom or a resource room; before, during, or after school); how often tutors met with students and for how long; whether the intervention lasted a week, a month, a semester, or a year; and whether the students received some incentive to take part in the tutoring and to stay involved.

The JARS module on experimental manipulations also asks whether the researcher (a) randomly assigned participants to conditions or (b) assigned participants to conditions using another type of procedure; depending on the assignment procedures, the researcher asks additional questions about how the assignment was accomplished.

Of course, this is just a sampling of the information requested in the JARS. As you look at the full JARS, it might seem daunting to provide all the information. But really, everything asked for should be known about a study. The JARS helps researchers remember what’s important about their study and ensures their study remains a valuable contribution to the psychology literature.

December 10, 2010

Journal Article Reporting Standards: Why Are They Needed?

HCooper 3-1-09

by Harris Cooper, PhD


Harris Cooper, PhD, was chair of the APA Journal Article Reporting Standards Working Group. He also served on the committee that revised the APA Publication Manual.

 With the holidays around the corner, nothing frustrates us more than incomplete assembly instructions for that bicycle or bookshelf. We fume over the instructions that are unclear or the list of materials that don’t quite match up with the material provided. There seems to be a screw missing. What is this piece for? Does the shelf go in before or after tightening the screws?

88349067-sm

In many ways, a psychology research report is like assembly instructions. Without a complete list of the materials and a clear description of the assembly steps, it is impossible for others to understand what we did and what to do to repeat our experiment, if they so desire.

Recently, more people have become interested in what the psychological research says. But, with increased influence comes increased responsibility, and increased scrutiny. And, there has been a growing sense that the instructions in our research reports often do not serve us well.

A desire for “evidence-based” practice is widespread in public health, social services, and education. Before funding a program to, say, reduce drug abuse, improve academic achievement, or assist veterans with posttraumatic stress disorder, the people who control the purse strings want “solid evidence” the program works. Solid evidence means that the studies that purport to evaluate programs and interventions allow confident conclusions about program effectiveness. And, to make this judgment, the research write-ups need to include clear instructions and an accurate list of materials.  How else will they know whether studies’ results are to be believed?

In addition to this need for easy replication is a desire for uniformity in discussing results.  The amount of psychological research is growing rapidly. When researchers summarize studies, be they about basic topics—such as the influences on memory or the development of morality through the life span—or applied topics, they need to have good descriptions of what was done. Like assembly instructions, these descriptions are used to piece together past research into coherent pictures, to help resolve conflicts in research results, and to identify questions yet to be studied. If the research description is incomplete, it is like assembly instructions that result in a bicycle that we can’t ride or a bookshelf that will collapse.

Not surprising then, greater emphasis today is placed on the reporting of research. So, in preparation for the sixth edition of the Publication Manual, APA formed a working group to look into the issue. As a result, the Publication Manual now recommends that Journal Article Reporting Standards (or the JARS) be followed that summarize the information editors, reviewers, and readers will expect to see in research reports. APA has just released a book I authored to help writers understand and implement the new standards, titled Reporting Research in Psychology: How to Meet Journal Article Reporting Standards. So, how do the JARS work? Find out in next week’s post.

 

July 01, 2010

A Post About Nothing

Timothy McAdoo by Timothy McAdoo

Today I want to zero in on a special topic. This is not just an empty set of words: Let’s de-cipher another APA Style point!

The zero before a decimal point is known as a leading zero. Have you noticed that sometimes this zero is used in decimal values and sometimes it is not?

APA Style has a very simple guideline for leading zeros:

  • If a value has the potential to exceed 1.0, use the leading zero.
  • If a value can never exceed 1.0, do not use the leading zero.

Thus, because most units of measure have the potential to exceed 1.0, the leading zero is frequently needed. A value over 1.0 does not need to actually appear in the text. Here are just a few examples:

…was 0.75 in. tall by 0.95 in. wide.

Participants viewed, on average, less than 0.65 hr of the footage.

…had means of 1.01, 2.21, and 0.95, respectively.

…had 95% CIs [0.62, 1.12], [-2.44, 4.30], and [-3.19, -2.39], respectively.


There are some values that by definition can never exceed 1.0. The omission of the leading zero is a visual indicator of this restricted range. The most common cases are p values and correlations:

…was significant (p < .01).

…was significant (p = .001).

…was shown to be highly correlated (r = .71).


A consistent presentation of statistical values, both within a paper and across published articles, provides a visual symmetry that can help readers focus on content over form.

I hope this zippy post has helped nullify any confusion. If you’re still drawing a blank, you can also find this guideline and additional examples on page 113 of the Publication Manual of the American Psychological Association (6th ed.).

June 17, 2010

Formatting Statistics: Using Brackets

Timothy McAdoo by Timothy McAdoo

In the previous post, we discussed how to use parentheses and commas with statistics.

Today, we highlight an exception to this guideline. There are two cases when brackets are the preferred choice in APA Style.

First, when parenthetical text also includes a statistic, brackets should be used.

(See Figure 3 for the results from the control group [n = 8]; compare with results from the Pink Floyd listening group [n = 23] and the Beatles listening group [n = 41].)

Second, for clarity, APA Style recommends that confidence intervals be reported with brackets around the upper and lower limits (as outlined on page 117):

95% CI [5.62, 8.31]

In the context of a sentence this might look like the following:

Participants who heard one Dresden Dolls song on repeat for 180 min reported no less anxiety than those who heard one Mozart movement on repeat for 180 min, R2 = .22, F(1, 32) = 7.33, p = .003, 95% CI [0.11, 1.23]. Our hypothesis that the genre difference would influence anxiety levels was rejected.

Also, as noted in the Publication Manual of the American Psychological Association (6th ed.) on page 117, every report of a confidence interval must clearly state the level of confidence. Multiple confidence intervals would appear as follows. Notice the plural version of the CI abbreviation:

... 95% CIs [5.62, 8.31], [-2.43, 4.31], and [-4.29, -3.11], respectively.

I hope these examples are helpful.  What other questions about formatting statistics do you have? Let us know in the comments.

June 11, 2010

Formatting Statistics: Using Parentheses

Timothy McAdoo by Timothy McAdoo

The Publication Manual (6th ed.) presents guidelines for formatting statistical and mathematical copy. As noted on page 116 of the Manual, these guidelines reflect "standards of content and form agreed on in the field" and are designed to enhance clear communication.

The most basic guidelines are that statistics should be italicized and shown in parentheses:

In Experiment 1, participants listening to The Smiths were no more likely to dance energetically than were those listening to The Cure (p = .24).


However, when a statistic includes its own parenthetical value (e.g., degrees of freedom that appear with t or F values), you’ll need to separate the statistics from the text with commas. Nested parentheses should be avoided in APA Style.

Consider the next two examples:

In Experiment 2, participants listening to Lady Gaga were more likely to “just dance” than were students listening to The Smiths, t(177) = 3.51, p < .001.

In Experiment 3, participants’ biofeedback indicated a significant impact of listening to up-tempo La Roux melodies, F(1, 144), p < .001, and also a significant impact of listening to melancholy selections from The Decemberists, F(1, 144), p < .001.


Notice how the commas separate the string of statistics from the text, just as they would separate other types of independent clauses.

When you have multiple groups of statistics in a series, use semicolons to separate:

In Experiment 4, an impact was demonstrated for genre, R2 = .31, F(2, 13) = 3.13, p < .001; recording date, R2 = .11, F(2, 13) = 1.53, p < .001; and tempo, R2 = .17, F(2, 13) = 2.33, p < .001.


As a side note, notice that there should be no space between a statistic's symbol and its parenthetical information (in this case the degrees of freedom): that is, F(1, 144) not F (1, 144).

Of course, this just scratches the surface of the potential for reporting statistics. Pages 116–124 of the Manual provide much more detail. What questions do you have? Let us know in the comments.

February 18, 2010

Lists, Part 2: Commas and Semicolons

Timothy McAdoo by Timothy McAdoo

This is the second in a six-part series about lists in APA Style. Today I’ll provide examples of serial commas and semicolons.


Commas

The most basic type of list appears in the running text of a sentence, with each item separated by a comma. All lists in APA Style should include a serial comma—the final comma before the conjunction.

For example,

Each child was given a plush toy, a building block, and a rubber ball.


Semicolons

This gets more complex when an item or multiple items in your list already have commas. In these cases, separate the items with semicolons:

Each child was seated at a separate station and given the following plush toy or toys: an elephant, which all children saw in the previous experiment; a kangaroo, which only half of the children saw in the previous experiment; or both the elephant and the kangaroo.


In the next example the same principle is applied to a series that includes statistics. Proper and consistent use of commas and semicolons clarifies the grouping of each set of statistics:

The results of Experiment 1 showed a similarity across groups: Group A, t(177) = 3.01, p < .001; Group B, t(173) = 2.31, p < .001; and Group C, t(155) = 3.11, p < .001.


More examples of commas and semicolons within lists can be found in the Publication Manual of the American Psychological Association on pages 63–65 (3.04 Seriation), on page 88 (4.03 Comma), and on pages 89–90 (4.04 Semicolon).

More to Come

In Part 3, we’ll look at how to use lowercase letters to identify elements of a list in APA Style.

Lists, Part 1  |  Lists, Part 2  |  Lists, Part 3

Search the APA Style Blog


ABOUT THE BLOG

My Photo


About Us

Blog Guidelines

APA Style FAQs

Archives


rss Follow us on Twitter

American Psychological Association APA Style Blog

Twitter Updates