5  Tips and Advice

5.1 Things to Avoid at all Cost

Disseminating research results, rather in a paper or a poster presentation, can be challenging, especially if this is your first time doing so. The following is a brief list of things to avoid at all cost when writing about or presenting on this project:

  • Avoid any reference to this class
    • “Our group did this…”
    • “For this research project we…”
    • “In this class…”
  • Avoid excessive first person language
    • The method allows for some more “we’s” and “I’s”
  • Avoid a history of how you “got here”
    • Write “the article that makes the most sense now that you have seen the results” (Bem, 2003) rather than the one you thought you were going to write on day 1.
  • Avoid willy-nilly summaries of statistical tests. Don’t invent your own ways to describe statistical results…
    • There is a formal way to present statistical results (see the section on this topic in this project manual)
  • Avoid using the word “correlation” to mean anything other than an actual statistical correlation such as Pearson’s Correlation (the little “r”)
  • Never say that your results proved something
  • Don’t define terms by quoting or citing a dictionary. This practice is surprisingly common among college students. Words that show up in a dictionary don’t need a citation. Other phrases and terms that are studied by academics should be defined by describing them, citing peer reviewed research to bolster claims, etc.
  • Avoid colloquial, every day conversational language. Your documents should be written in a formal tone.
    • This doesn’t mean you should use super-sciency language either. Use professional language, using “fancy” terminology appropriately
  • Avoid writing a Sherlock Holmes novel. You are not writing a mystery novel. Tell your reader exactly what’s coming rather than hinting at something special around the corner. Scientific writing should be direct and clear (that doesn’t mean you cannot have a nice narrative that weaves everything together).
  • Don’t let numbers/results/statistics speak for themselves. They never do!
    • Tell your reader what happened in conceptual terms, and then show the statistical result to demonstrate that the effect you found was significant. The statistical result is not the interesting thing. It is necessary to show that your result is significant, but the result itself is the thing the reader actually cares about. So talk about the result, and simply use the statistical result to buttress your claims.

5.2 Survey Creation Todo List

Many projects will at least partially involve the creation of a survey. The following guidelines provide a brief outline of what you will need to do to create and deploy a survey for this project.

Important

When using Qualtrics for your project, make sure you always visit Capital’s Qualtrics service by using this link: https://capital.qualtrics.com/. If you search Google for Qualtrics, you will get linked to www.qualtrics.com, which will not work. After creating an account on Capital’s Qualtrics, you still will NOT have an account at www.qualtrics.com. So be sure to bookmark or remember the correct Qualtics link.

If you have trouble signing in, due to (possibly) forgetting your Capital password, do not click the password reset link. This will not actually let you reset your password, but it will temporarily block you from signing in to Qualtrics.

  1. Activate your Qualtrics account. This is a multi-step process. You must complete all of the following steps before your Qualtrics will be fully activated.
    1. Visit the following link.
    2. If this is your first time using Qualtrics, log in with your usual Capital credentials (leave off the “@capital.edu” from your user name).
    3. On the next page, you will be asked if you’ve ever used Qualtrics before. Even if you have a preexisting account with another organization, say this is your first time using Qualtrics. A new account will be automatically generated for you.
    4. Complete the user agreement.
    5. Once you have completed the user agreement, your account will be activated, typically within 48-72 hours.
    6. If you run into any issues, email qualtrics@capital.edu.
  2. You will need to write a brief informed consent to your survey that tells the participant what is expected of them, along with any risks or requirements. This should be the first thing the participant sees when they start your survey. Set the survey so that they have to choose “agree” before they can continue with the survey. If you use the participant pool to recruit participants, you can also use this text as the description when creating the study. Things to include in this brief paragraph:
    • You must be at least 18 years of age to participate
    • Include any other requirements, such as you must be a user of SnapChat, must use an iPhone, etc.
    • You freely choose to enroll in this study. You may refuse to enroll instead, and may leave the study at any time without losing any benefits to which you would otherwise be entitled. The researchers may choose to end your participation early at any time.
    • If your study is completely benign, then you should included a statement to the effect of, “There are no known risks for participating in this study.” Given that your study will not have gone through IRB approval, it’s important for this to be true. In the event that you’ve been given special permission to include content that is potentially provocative, it will be important to appropriately warn potential participants in place of the previous statement.
    • You will not receive any benefits for participating in this study aside from research participation credits
    • Your responses are anonymous and will not be linked to your real identity
    • A brief description of what the participant will have to do. Note that your goal is not to tell the participant what your research questions or hypotheses are. Rather, the goal is to inform them of what they should expect while completing your study. This is to ensure that they are comfortable with what you plan to ask. For instance, if you plan to ask questions about daily stresses/anxiety, you should let them know that they will have to share this potentially personal information.
  3. Have one group member create your survey in Qualtrics
    • Make the first question in your survey your informed consent (see above). The “question” text should be the informed consent, and you should provide two response options: I consent to participate and I do not consent to participate.
    • Use skip logic to control the survey’s behavior. Specifically, if the participant chooses not to agree with your informed consent, you should have the survey jump to the end (so that they are not exposed to the remaining questions in the survey).
  4. The group member that created the survey must then add every other group member as a collaborator in Qualtrics. This can be done navigating to your main Qualtrics dashboard (the first page you see when signing in, or by clicking the “XM” icon near the top-left of any Qualtrics page). From this main dashboard, click the three dots next to your newly created survey and select “Collaborate.” You can then search for each group member by name, adding them to your survey. After completing this, each group member should be able to view and edit the survey.
Tip

If you cannot add survey collaborators because the menu option is unavailable, you do not have a fully activated Qualtrics accounts yet.

If you can add collaborators, but one of your group members doesn’t show up when you search for their name, then they do not have a fully activated Qualtrics account yet.

In either case, a full account can be achieved by completing the user agreement mentioned above.

5.3 Reporting Statistics in APA Format

Important

Avoid willy-nilly summaries of statistical tests. Don’t invent your own ways to describe statistical results…

When writing about the results of your statistical analyses, you are expected to use an appropriate APA-style format. What follows below are examples of how to properly report findings using some of the most common statistical tests.

5.3.1 Chi-squared Results

Results to report:

  • X2
  • degrees of freedom
  • sample size (N)
  • significance p

Example:

  • A significant difference was found between groups. Reddit users were found to be more interesting to others than Instagram users, X2(1, N = 112) = 7.8, p = .002.

5.3.2 Correlation Results

Results to report:

  • correlation coefficient r
  • significance p

Examples:

  • Instagram use and depressive symptoms were found to be positively correlation, r = .53, p = .023.
  • No relationship was found between alcohol consumption and fourth year GPA, r = .07, p = .84.
Note

Notice the use of italics when reporting statistics.

5.3.3 t-test Results

Results to report:

  • mean (M) of each group
  • standard deviation (SD) of each group
  • t value(t)
  • degrees of freedom
  • significance p

Examples:

  • College graduates (M = 5.34, SD = .50) reported significantly higher levels of happiness than non-graduates (M = 3.45, SD = .37), t(1) = 6.31, p = .001.
  • TikTok users (M = 4.21, SD = .47) and Instagram users (M = 4.32, SD = .47) did not differ significantly on the factor of “fear of missing out,” t(1) = 1.27, p =.76.

5.3.4 ANOVA Results

Results to report:

  • mean (M) of each group
  • standard deviation (SD) of each group
  • F value (F)
  • degrees of freedom
  • significance p

Examples:

  • The main effect of gender was not significant, F(2, 97) = 1.27, p = .12. Male, female, and non-binary participants did not differ on the reported recreational drug consumption (see Table 1 for means).
  • A main effect of drowsiness was found, F(2, 67) = 14.63, p = .017. Participants reported significantly more sleepiness post-experiment (M = 6.00, SD = 0.28) than either before (M = 3.76, SD = .68) or during the experiment (M = 2.17, SD = 1.06).

5.3.5 Regression Results

Results to report:

  • R2
  • F value (F)
  • degrees of freedom
  • significance p

Example:

  • Multiple regression was used to determine if personality traits significantly predicted participants’ ratings of aggression. The results of the regression indicated the two predictors explained 35.8% of the variance (R2=.38, F(2,55) = 5.56, p = .02). It was found that extraversion significantly predicted aggressive tendencies (β1 = .56, p = .001), as did agreeableness (β2 = -.36, p = .012).