Materials

Participants saw one of three dominant messages and one of three non-dominant messages:

dom_1: By now you know the task at hand. If you don’t complete it and do it well, you will not get the full bonus.
dom_2: All right, it’s go time. If you want to keep your entire bonus, you’re going to need to jump in and nail this task.
dom_3: Ok, get in there and do your absolute best across all rounds. If you don’t perform well, I will take away some or all of your bonus.

nondom_1: I appreciate your hard work. Just look at the shapes carefully and try to get as many as you can right.
nondom_2: Please complete this task to the best of your ability. Thanks!
nondom_3: Your job in this task is to select the shapes that match the description. Make sure you look at them carefully.

Measures

Message chosen

Which message did people select?

message_type n
dom 19
nondom 77

hmm. Not as good as the earlier pilot. But maybe this is ok. Let’s see which dominant messages ones are chosen:

message n
dom_1 6
dom_2 10
dom_3 3
nondom_1 28
nondom_2 27
nondom_3 22

Predictions: Compliance

Do task

Will the employee do the task?

message prediction n
dom do-task 58
dom skip-task 38
nondom do-task 92
nondom skip-task 4

Performance

If they do the task, how many points will they score?

Predictions: Relationship

Another task: With manager

How much money will it take to get them to do another task with this manager?

Another task: Random

How much money will it take to get them to do another task with someone at random?

Another task: Difference

Price of doing the task with the current manager MINUS price of doing the task with someone at random.

we like this.

Attitude

What will be the impact of this message on your employee’s attitude towards you?

Projection

If you were an employee and you received the message above from your manager, what would be its impact on your attitudes towards your manager?

Punish decision

For each level of their performance below, please indicate what bonus you would like them to receive. You can select between 0 and 200 cents for each. 200 cents is the full $2.00.

Competitive Worldview

1 = Strongly Disagree to 7 = Strongly Agree

1. It’s a dog-eat-dog world where you have to be ruthless at times
2. Life is not governed by the “survival of the fittest.” We should let compassion and moral laws be our guide [R]
3. There is really no such thing as “right” and “wrong.” It all boils down to what you can get away with
4. One of the most useful skills a person should develop is how to look someone straight in the eye and lie convincingly
5. It is better to be loved than to be feared [R]
6. My knowledge and experience tell me that the social world we live in is basically a competitive “jungle” in which the fittest survive and succeed, in which power, wealth, and winning are everything, and might is right
7. Do unto others as you would have them do unto you, and never do anything unfair to someone else [R]
8. Basically people are objects to be quietly and coolly manipulated for one’s own benefit
9. Honesty is the best policy in all cases [R]
10. One should give others the benefit of the doubt. Most people are trustworthy if you have faith in them [R]

Cronbach’s alpha = 0.8

Correlations

is_dom: dummy-coded message chosen (0 = nondom; 1 = dom).
man: price to get someone to do the new task with the current manager.
rand: price to get someone to do the new task with someone at random.
diff: price of man MINUS price of rand.
attitude: predicted impact of the message on employee’s attitude towards the manager.
proj: impact of the manager on the participant if they participant was the employee.
comp: score the employee will get in the task if sent this message.

Analysis

CWV -> predicted difference in price

(#tab:unnamed-chunk-15)
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept 0.55 [-0.06, 1.17] 1.78 94 .078
CWV -0.07 [-0.28, 0.14] -0.64 94 .526

CWV -> choosing dominant message

(#tab:unnamed-chunk-16)
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept -0.14 [-0.40, 0.12] -1.08 94 .283
CWV 0.12 [0.03, 0.21] 2.72 94 .008

predicted difference -> choosing dominant message

(#tab:unnamed-chunk-17)
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept 0.23 [0.15, 0.32] 5.43 94 < .001
Dom diff -0.10 [-0.19, -0.01] -2.26 94 .026

Mediation model: CWV -> predicted price difference -> choosing dominant message

a = -0.07 (p = 0.526)
b = -0.09 (p = 0.034)
direct = 0.12 (p = 0.008)
indirect = 0.12 (p = 0.01)

Mediation model: CWV -> predicted impact on attitude -> choosing dominant message

a = 0.34 (p = 0.073)
b = 0.11 (p = 0)
direct = 0.12 (p = 0.008)
indirect = 0.09 (p = 0.039)

Compare to ground truth

Let’s see some mean scores from our employees. First, just a refresher, these are the messages they have seen:

dom_1: By now you know the task at hand. If you don’t complete it and do it well, you will not get the full bonus.
dom_2: All right, it’s go time. If you want to keep your entire bonus, you’re going to need to jump in and nail this task.
dom_3: Ok, get in there and do your absolute best across all rounds. If you don’t perform well, I will take away some or all of your bonus.

nondom_1: I appreciate your hard work. Just look at the shapes carefully and try to get as many as you can right.
nondom_2: Please complete this task to the best of your ability. Thanks!
nondom_3: Your job in this task is to select the shapes that match the description. Make sure you look at them carefully.

message points again_man again_rand again_diff attitude
dom_1 31.70588 1.875000 2.125000 -0.2500000 3.411765
dom_2 32.87500 2.312500 2.281250 0.0312500 4.687500
dom_3 27.52941 2.470588 2.470588 0.0000000 3.294118
nondom_1 29.41667 2.416667 2.250000 0.1666667 5.833333
nondom_2 24.95000 2.400000 2.525000 -0.1250000 4.700000
nondom_3 36.11111 2.055556 2.333333 -0.2777778 4.222222

Again_diff

The difference score between doing it again with the manager and doing it again with someone at random. I’ll show the relationship between CWV and that predicted score, for managers. And I’ll overlay the means from the employees. All of this will be per message.

We only care about dominant messages for now. So I’ll start with that.

Ok, seeing no relationship for dom_1 and dom_2. But for dom_3, we’re seeing the relationship in the expected direction (high CWV don’t expect a premium for doing the second task with the same managers), but it’s driven by low CWVers being overly pessimistic. Ok cool.

Let’s see non-dominant messages while we’re at it.

Attitude

Dominant messages

Looks a little different here. For this one, low CWVers are off in dom_2 and dom_3, and both are off in dom_1.

Non-dom messages

Points

Dominant messages

Non-dom messages

Additional analysis

relationship or compliance?

Model 1

Predicted attitude AND predicted points -> message choice

(#tab:unnamed-chunk-27)
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept -0.27 [-0.49, -0.06] -2.50 93 .014
Attitude dom 0.10 [0.05, 0.15] 4.27 93 < .001
Comp dom 0.00 [0.00, 0.01] 1.57 93 .120

Wow. That’s pretty. Attitude is driving this effect.

Model 2

Predicted price premium AND predicted points -> message choice

(#tab:unnamed-chunk-28)
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept -0.08 [-0.30, 0.15] -0.66 93 .512
Dom diff -0.09 [-0.17, 0.00] -2.04 93 .045
Comp dom 0.01 [0.00, 0.01] 2.91 93 .004

ok, compliance is doing most of the work here. But still, there’s some extra variance explained by the predicted premium of participating in a second task with the manager.