GB730 Unit 1 Discussion Board Post Example

Course	GB730 Applied Business Research Design and Methodology
Unit	Unit 1
Paper type	Discussion board post
Length	About 580 words, 3 pages double-spaced
Format	APA 7 discussion post
School	Purdue University Global
Program	MBA
Updated	October 2026

Free sample paper for GB730 Unit 1

Overridden Nights Earned More, So Overrides Work? A Unit 1 Discussion Post on Claims Without Random Assignment

GB730 · Unit 1 Discussion Post

Posted by [Student Name]

The hotel company, its figures and its people are composite and illustrative; the sources cited are real.

What this part is doingStating the executive's reading of the gap in the lead shows the post is about the inference, not the number.

Initial Post

I work as a senior revenue analyst at a composite hotel management company that runs 22 select-service and full-service hotels for their owners across the Midwest and Southeast. In 2022 the company installed a revenue management system that recommends a room rate for every hotel and every future night, updated daily as bookings come in. Each hotel's revenue manager can accept the recommendation or override it. At a meeting last quarter, our senior vice president of revenue showed a slide: on nights where managers overrode the system, revenue per available room averaged 118 dollars; on nights where they accepted it, 104 dollars. "Our people beat the machine," she said, and proposed giving managers more authority to override.

The difficulty is that overridden nights differ from accepted nights before anyone overrides anything. Managers override most often when they know something the system does not: a convention booked late, a concert announced after the forecast ran, a competitor's hotel closed for renovation. Those are nights when demand is already unusually high. A night with a sold-out concert will earn more revenue whether or not the manager touches the rate. The comparison on the slide therefore mixes the effect of the override with the effect of the conditions that prompted it. Shadish et al. (2002) call this a selection threat: the groups being compared were formed by a process related to the outcome.

What could a study of these nights defend without random assignment? It could estimate how overrides are associated with revenue after adjusting for measured indicators of demand, such as bookings already on hand two weeks out, local events and competitors' rates. It could describe when managers override and when they do not. It could explore, through interviews, what managers know that the system does not. It could not, on its own, show that overriding causes higher revenue, because unmeasured knowledge, the very reason managers override, would remain in the comparison.

The nights a manager overrides are the nights a manager knows something, and that knowledge earns money with or without the override.

The obvious solution, randomly assigning some nights to forced acceptance of the system's rate, is not available. Managers' bonuses depend on revenue, owners' management agreements give the company a duty to maximize revenue for each hotel, and no owner would agree to a lottery that might cost money on high-demand nights. Dietvorst et al. (2015) found that people tend to distrust algorithms after seeing them err, which suggests the slide may also say something about managers' confidence rather than about outcomes.

My question for classmates: in your own setting, what comparison is being read as proof when the groups were formed by the very conditions that drive the outcome?

Reply to Marcus

Marcus, your plan to compare call center wait times before and after your new scheduling software has a different problem. Comparing the months before and after the software arrived mixes its effect with whatever else moved in that window, such as seasonal call volume or a new product launch. Could you add a comparison group, for example a second center that adopted later, so the change over the same months can be compared?

What this part is doingThe reply names the specific threat in the classmate's design rather than giving general encouragement.

References

Dietvorst, B. J., Simmons, J. P., & Massey, C. (2015). Algorithm aversion: People erroneously avoid algorithms after seeing them err. Journal of Experimental Psychology: General, 144(1), 114-126. https://doi.org/10.1037/xge0000033

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.

Reading the GB 730 Unit 1 instructions

The opening GB730 board wants a comparison from your own workplace that someone has read as proof of cause. Describe the setting and the people making the decision before giving any figures. Quote the comparison as it was presented, with its numbers. Then explain how the two groups being compared came to differ before the decision was made. Name the threat to internal validity from a methods text rather than calling it bias in general terms. List what a study without random assignment could still defend, and say plainly what it could not. Explain why an experiment is not available in your setting. Finish by asking classmates where a comparison in their own work was formed by the conditions that drive its outcome. The reply then names the specific threat in one classmate's design and proposes a fix.

Inside the GB730 Unit 1 example

The post introduces the hotel company, its 2022 pricing system and the senior vice president's slide before any analysis. It then shows why overridden nights are different nights: a late convention, a concert, a competitor closed for renovation. Selection is named from Shadish and colleagues. Four things the study could defend follow, from adjusted associations to interviews about what managers know, with the causal claim it cannot make stated after them. The highlighted sentence puts the whole problem in one line. A paragraph explains why owners and bonus plans rule out randomizing nights, and Dietvorst's 2015 study adds a second reading of the slide. The reply to Marcus spots history as the threat in a before and after comparison of call center software and suggests a later adopting center as a comparison group.

Where the points sit in the GB730 Unit 1 rubric

In GB730 the opening post earns most of its points for reasoning about design rather than for the topic chosen. Instructors look first at whether the comparison is quoted exactly, since vague paraphrase hides the inference being tested. Explaining how the groups were formed, and naming the specific threat that formation creates, carries the most weight on this board. A list of defensible claims shows the student can still build a study after admitting the limit. The section on why randomization is impossible must rest on facts of the setting, such as contracts or incentives. Citing the threat from its original methods source counts more than citing a summary. The reply is graded on whether it identifies a different threat accurately and offers a design change, not on encouragement.

Common GB730 Unit 1 mistakes, and how to avoid them

A frequent weakness in the GB730 opening post is treating any difference between groups as a causal effect and then criticizing the size of the number instead of its source. Some students call the problem bias without naming selection, history or maturation, which leaves the instructor unsure they know the distinction. Others propose an experiment without checking whether their organization could ever run one. The list of what the study can defend is often skipped, so the post ends in pessimism rather than a plan. Replies drift into praise for a classmate's topic. Quote the comparison, explain how the groups were made, name the threat, state the defensible claims and the impossible one, and give the classmate a concrete fix. For a version built on your own workplace comparison, send us the claim you want to test.

Get a GB730 Unit 1 example written to your instructions

Paste the question posted for the Unit 1 discussion, say which business setting or research interest is on your mind and add the rubric. Inside 24-48h a custom post ranks the claims a nonrandomized study could defend and adds a reply that tests a classmate's design against one named threat. Your first one costs nothing. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.

More GB730 papers and MBA samples

GB730 Unit 1 questions, answered

Does GB730 expect a research topic in Unit 1?

Many sections want a direction rather than a settled topic. A sentence naming the setting and the question that interests you is usually enough, since the problem and purpose work in following units will reshape it. What the first post needs more than a topic is evidence that you can read a design critically and name its weak point.

Which methods texts suit the Unit 1 post?

Shadish, Cook and Campbell on experimental and quasi-experimental designs remains the standard source for threats to validity, and Creswell's research design texts serve as the course text in many sections. Citing the specific threat, such as selection or history, from the original source reads far better than a general reference to bias.

Is a quasi-experimental design weaker than an experiment?

For causal claims, generally yes, though the gap depends on the design. A well-built comparison with a credible control group can support strong claims, while a simple before-and-after comparison supports weak ones. Applied doctoral work often relies on quasi-experimental or correlational designs because organizations rarely allow randomization, so the task becomes stating claims at the right strength.

Where can I find a free GB730 Unit 1 sample paper?

This page carries a full GB730 Unit 1 discussion post on a hotel company's override slide, naming selection as the threat, listing what a nonrandomized study could defend and closing with a reply to a classmate.

What is a selection threat in a nonrandomized study?

It arises when the groups being compared were formed by a process linked to the outcome. Here managers override on nights they expect strong demand, so overridden nights would earn more even with no override at all.