Testing Hypotheses, Part 1: Overview

Loading...

来自 University of California, Davis 的课程

Quantitative Research

29 个评分

University of California, Davis

Quantitative Research

29 个评分

课程 3（共 4 门，Specialization Market Research）

In this course, you will obtain some insights about marketing to help determine whether there is an opportunity that actually exists in the marketplace and whether it is valuable and actionable for your organization or client. Week 1: Assess methods available for creating quantitative surveys, along with their advantages and disadvantages. Identify the type of questions that should be asked and avoid unambiguous survey questions. Week 2: Design, test, and implement a survey by identifying the target audience and maximizing response rates. You will have an opportunity to use Qualtrics, a survey software tool, to launch your own survey. Week 3: Analyze statistical models that can be applied to your marketing data, so that you can make data-driven decisions about your marketing mix. Week 4: Predict most likely outcomes from the marketing decisions and match the type of analysis needed for your business problem. Take Quantitative Research as a standalone course or as part of the Market Research Specialization. You should have equivalent experience to completing the second course in this specialization, Qualitative Research, before taking this course. By completing the third class in the Specialization, you will gain the skills needed to succeed in the full program.

从本节课中

Analyzing a Survey

In this module, you will be able to review your survey data and look for any errors. You will be able to analyze your survey conducting descriptive and inferential analysis techniques to test your hypotheses. You will be able to conduct association analysis and causal analysis with your data. Please note that this module features a fair amount of advanced statistics related math. Throughout the module, optional activities setup as practice quizzes have been provided to help reinforce what you're learning.

Analyzing a Survey Introduction1:09

Reviewing Your Data and Looking for Errors4:47

The Importance of Scales5:50

Descriptive Analysis: Nominal Data6:14

Descriptive Analysis: Ordinal and Interval Data5:21

Inferential Analysis: Confidence Interval6:27

Testing Hypotheses, Part 1: Overview3:27

Testing Hypotheses, Part 2: An Example Application6:40

Association Analysis: Comparing Groups of People4:37

Association Analysis: Chi-Square Test6:02

Association Analysis: Correlation Analysis2:01

Causal Analysis6:15

与讲师见面

Susan Berman
President, Impact Research
UC Davis Division of Continuing and Professional Education
Olivier Rubel, PhD
Associate Professor, Graduate School of Management
UC Davis

One of the goals of marketing research is to

formulate and test hypotheses that are relevant for marketing decisions.

These questions may be, for example,

are people different from one another?

Is there willingness to pay for a product $305 or is it larger?

So hypothesis testing is a statistical procedure

that gets you to accept or reject a claim based on two statistical tests,

namely Z-Test and T-Test.

Similar to inferential analysis, it is important that

the sample is representative of the population of interest.

Specifically what you want to do is to

infer something from a larger group, the population.

However, you cannot survey everyone.

So you take a sample of this larger group.

To formulate hypotheses that amenable to testing,

we need to define what is called the null and

alternative hypothesis which are usually denoted H_sub_0 and H_sub_a respectively.

The alternative hypothesis is what we want to

test whereas the null hypothesis is the default assumption.

The null hypothesis is usually formulated as an equality about the quantity of interest

whereas alternative hypothesis can be formulated with a different sign or an inequality.

The objective of hypothesis testing is to see whether there is enough information in

the data to reject the null hypothesis and accept the alternative hypothesis.

Usually and ideally, hypotheses should

be formulated before you design the survey and collect the data.

And hypothesis testing should inform questionnaire and sample design.

Here I'm assuming that these hypotheses have been made already and that now,

you want to move forward and analyze the results to see if

there is enough support in the data to reject or accept your hypothesis.

So the first step is to formulate the hypotheses

based on what you think is true for the population.

Then, based on the statement,

based on the data, based on the scale used,

and based on the confidence level you want

to determine something called the sample statistic.

And you want to compare it to

the positive parameter and decide whether or

not the data support the original hypothesis.

Suppose that you want to test for means, means being equal,

means being different in the population compared to

the sample size or means being larger or smaller than a specific value.

Imagine that you have interval data on some means.

To make some hypotheses using your sample,

you can use one of two test: something called Z-Test and something called the T-Test.

They are about the same in the sense that they're designed to test hypotheses or means.

However it is important that we distinguish between the two of them.

Z-Test is appropriate when the samples,

where the population standard deviation is assumed to be known whereas

a T-Test is appropriate when both the

mean and the standard deviation of the populations are unknown.

So we'll focus on the Z-Test for simplicity.

The T-Test and the Z-Test,

however, are not too different from one another.

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2018 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取