'Meissner 2016 (ExpEcon) - partial replication of adapted design'
(AsPredicted #77627)

---

---

1) Have any data been collected for this study already?
No, no data have been collected for this study yet.

2) What's the main question being asked or hypothesis being tested in this study?
This study is a partial replication of a modified design version from Thomas Meissner (2016, Experimental Economics) - "Intertemporal consumption and debt aversion: an experimental study". Its main question re-visited here is whether people treat savings systematically differently than debt with respect to consumption smoothing.

3) Describe the key dependent variable(s) specifying how they will be measured.
Deviation (absolute or directional) from conditionally optimal consumption (m1 and m2), and unconditionally optimal period-consumption as a function of optimal wealth (m3) - as defined in the original paper.

4) How many and which conditions will participants be assigned to?
Participants will be randomly assigned to either 'Savings First' or 'Borrowing First' condition, aiming for equal-size cells (small cell size differences might arise from uneven subject attrition across treatment cells).

On Savings (Borrowing) First, subjects face a decreasing (increasing) income path per period. The expected lifetime earnings are the same in both treatments. After playing two rounds (i.e. experimental lives 1-2) in their initially assigned condition, subjects experience the other condition for the remaining two rounds (3-4).

5) Specify exactly which analyses you will conduct to examine the main question/hypothesis.
Following the original paper, we will use Mann-Whitney-U tests for between-subject comparisons from two treatment groups' smoothing measures (m1, m2 and m3) to check whether these measures differ across treatment groups.

6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations.
We will exclude subjects that (1) take less than 60 seconds to read through the instructions; (2) answer more than one (out of six) wrong questions in a multiple-choice quiz after a trial round in a first attempt; (3) answer exactly one wrong quiz question after a first attempt, and still get it wrong on a second attempt - after being given a hint from the instructions; (4) take more than 60 minutes to finish all experimental tasks after completing the quiz; or (5) drop-out or abandon the experiment before its last screen.

7) How many observations will be collected or what will determine sample size?
No need to justify decision, but be precise about exactly how the number will be determined.
An active entry link will be kept at Bilendi (our panel recruitment contracted service) until it receives 464 backlinks of full successfully completed experiments, or until 01.Nov.2021 - whichever comes first. Subjects who are already in the completion pipeline when the entry link is shut down will be included in the final sample if they complete it timely as well. The pool from which we recruit comprises adult Germany-resident individuals pre-screened and subscribed to Bilendi's market research panel.

8) Anything else you would like to pre-register?
(e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?)
To adapt the author's original experiment to an online remote web environment using general population subjects, we reduced the length of each round to 10 periods, and the number of rounds to 4. We adjusted the expected values of the income path to range from 10 to 100 points. We rescale the implicit utility function parameters without changing its shape. We disallow negative spending (except when enforced automatically at the last period to assure that the remaining endowment after the last period is necessarily zero). We substitute Points for Taler and Eurocents for Points, respectively, converting the ECU Points directly to Eurocent-denominated rewards at each period. We also compensate subjects based on their sum of rewards accrued in one of the four lives, randomly chosen and revealed at the end of all tasks.

After their main experimental tasks, subjects will complete one-shot risk- and time-preference elicitation tasks. Results will be used in additional ANOVA models to control for effects from possibly preference heterogeneity across the two treatment groups. They will not be used for inference with respect to the main research question of the original paper.

---

Report a bug     |   Make a suggestion     |   Terms of use     |   Help     |   About