Paper III

1). We will continue to focus on the same reactance manipulation from study one for our first independent variable, but we will only keep the “High Controlling” versus “Low Controlling” levels of that independent variable (We will drop the “Neutral” condition. It often overlapped with the “Low Controlling” group in the study one data analyses, so keeping both the “Neutral” and “Low Controlling” conditions is needlessly repetitive. The High vs Low conditions provide a stronger manipulation, too, so we will retain both for Study Two. I’ll continue to call this the reactance condition. Thus, we will have the following reactance IV for our study design:

B). Low Controlling Language condition (The AI Policy uses low controlling language like “discouraged”, “probably avoid”, and “hope you choose”)

2). Now, consider our new independent variable again (agreement condition). Here we will note that a poll of students at FIU showed either high agreement for the proposed AI policy or low agreement with the policy.

A). For the high agreement condition, participants will be informed that a poll of students showed 86% agreement that it was a good policy for FIU to implement.

B). For the low agreement condition, participants will be informed that a poll of students showed 43% agreement that it was a goof policy for FIU to implement.

Nore: The % agreement is somewhat arbitrary here. We figured having double support in the high condition (86%) compared to the low condition (43%) adequately shows high versus low support for the proposed policy.

This creates a 2 (Reactance Condition: High Controlling Language versus Low Controlling Language) X 2 (Agreement Condition: High Agreement versus Low Agreement) factorial design. Thus, there will be four conditions:

Condition #1 – High Controlling Language and High Agreement

Condition #2 – High Controlling Language and Low Agreement

Condition #3 – Low Controlling Language and High Agreement

Condition #4 – Low Controlling Language and Low Agreement

As you begin writing your study two literature review for Paper III, keep this new “Agreement” independent variable in mind. You’ll need to find prior research that looks at Agreement and use that literature to help support or justify your study predictions. Fortunately, there are a lot of psychological areas that might be relevant here. Good keywords you might want to look for in PsycInfo are: “conformity”, “informational social influence”, “normative social influence”, “consensus”, “social norms”, “group cohesion”, etc.

Hypotheses:

For your hypothesis, you will need to focus on both main effects (the effect of each independent variable on its own) and an interaction (the influence of both independent variables interacting together). Each of your scaled dependent variables will need its own main effect and interaction hypotheses. Let me focus on one specific DV for now: “I intend to ignore the AI policy.”

1). Main Effect. IV = Reactance (High Controlling vs Low Controlling). DV = “I intend to ignore the AI policy”

Note #1: This replicates our study one prediction, though it lacks the “Neutral Language” condition. This main effect prediction thus only looks at the two levels of the Reactance independent variable (and it ONLY looks at Reactance).

“If participants read an AI policy that has low agreement from others, then they will more strongly agree that they intend to ignore the AI policy when compared to participants who read an AI policy that has high agreement from others.”

Note#1: The reasoning behind this prediction is that participants will most likely use other people as a cue about how they should react themselves (via both informational and normative social influence). If there is a high level of agreement, they might figure “I might as well support the policy (and follow it), too.” If agreement is low, they might figure “If others don’t like this policy, then I shouldn’t support it (and intend to follow it) either.”

Note #2: You will write your second literature review with this prediction in mind – find support to back it up! But again here, this main effect prediction ONLY looks at the “Agreement” independent variable. If your literature review research does not support this prediction, feel free to alter it, but you do need to justify why you think you might get your predicted outcome using prior studies in your second literature review.

Note #3: I admit I might be wrong here. Maybe people may not like the policy no matter how many others support it. In that case, participants may decide they will not follow the AI policy in any condition, regardless of whether 86% or 43% support it. My money is on the number of others mattering, though, but having “no differences” based on % provides us with a good null hypothesis.

3). Interaction, Reactance Condition (High Controlling Language vs. Low Controlling Language) X Agreement Condition (High Agreement vs. Low Agreement). DV = “I intend to ignore the AI policy”

“If participants read a high controlling language AI policy that has low agreement with others, then they will most strongly agree they intend to ignore the AI policy than all other conditions, with participants who read a low controlling language AI policy that has high agreement with others most strongly disagreeing that they intend to ignore the AI policy than all other conditions. Participants who read either a high controlling language AI policy that has high agreement with others or a low controlling language AI policy that has low agreement with others will not differ from each other, and they will fall between the other two conditions.”

Note #1: So here we do expect a significant interaction. In general, we expect the most “reactance” (as measured by intention to follow the policy) when participants see high controlling language AND there is low agreement that it is a good policy from others. In other words, people will rebel against a really strict policy, especially when others don’t like it! If you disagree with the prediction, that is fine. You can alter the prediction, but you do need to justify the predictions that you create.

Keep in mind that each dependent variable you plan to look at in your study two will need similar main effect and interaction hypotheses. I gave you ONE set of predictions for ONE of your dependent variables above. YOU must write your second set of predictions for a different dependent variable. Remember that it will have one main effect for the first IV (reactance), one main effect for the second IV (agreement), and an interaction of the two IVs (reactance X agreement). Thus your Paper Three will end with six predictions minimum (two main effects and an interaction for your first DV and two main effects and an interaction for your second DV). Be clear about what those predictions are in your second literature review so your paper builds toward them. Conclude the paper with those predictions. We will test them in Paper Four.

Good luck as you work on Paper III.