PCS - Planning and Conducting Surveys Overview

Math_APStatBanner.png Planning and Conducting Surveys Overview

Introduction

The word "survey" is used most often to describe a method of gathering information from a sample of individuals. This "sample" is usually just a fraction of the population being studied.   Surveys are being conducted on an ongoing basis:   radio stations conduct call-in surveys, websites ask for opinions and ratings, and telephone surveys poll voters' opinions during political elections.     If you spend any time at the shopping mall you may have participated in a survey conducted by a market research company.     Market researchers are hired by companies to gather consumer opinions about all different kinds of things like advertising methods, preferred color or taste of products, or even major design changes to automobiles.   The enormous quantity of this data requires computer analysis using commercial statistical packages, but the reports are written by humans who analyze the data and interpret it to give it meaning. 
A wide variety of businesses rely on surveys to help with decision making.   Major TV networks rely on surveys to tell them how many and what types of people are watching their programs.   You have probably heard about the Nielsen ratings. Auto manufacturers use surveys to find out how satisfied people are with their cars.   The U.S. Bureau of the Census conducts a survey each month to obtain information on employment and unemployment in the nation.   Local transportation authorities conduct surveys to acquire information on commuting and travel habits.   Magazine and trade journals use surveys to find out what their subscribers are reading. Surveys are conducted to ascertain who uses our national parks and other recreation facilities.   Everyday on the news you will hear reports about results of surveys or studies.   So you can see that using surveys is part of everyday life.   You have probably heard about Gallup Polls.   Gallup has built its reputation on delivering relevant, timely, and visionary research on what humans around the world think and feel.   You can find many reports about data they have analyzed at their website.   
The first step in conducting a survey is extremely important - who will be asked and how will they be selected?   Randomization is used to ensure a representative sample that is fair to all groups.   Bias can be introduced if any group is under-represented.   Careful question wording is essential so participants are clear about what is being asked.   All questions should be free of language that suggests one response over another. In this unit you will learn lots of new vocabulary associated with different random sampling strategies and all the potential pitfalls that can destroy the reliability of the data.   Let's get started!

Essential Questions

  • How do we design a survey that will result in the most reliable data?
  • WHO do we ask?
  • HOW do we ask them? What is bias?
  • How can biased responses be minimized? 
  • What strategies can be used for selecting an appropriate and representative sample?  

Key Terms

The following key terms will help you understand the content in this module.

Population- the ENTIRE group of individuals or instances about whom we want to learn

Sample- a representative subset (subgroup) of the population

Sampling variability- each sample will be different when drawing multiple samples at random

 Survey- a study that asks questions of a sample hoping to learn something about the population

Voluntary response sample- respondents can decide whether to respond or not...typically those with very strong opinions will respond more often making the sample less representative of the entire population

Convenience sample- subjects chosen based on availability and not on randomization

Simple random sample- is a sample where each element (or group of elements) has an equal chance of selection

Stratified random sample- population is divided into subgroups (strata) and random samples are drawn from each stratum

Systematic random sample- a sample drawn by using some 'system' typically every fifth subject or every 'nth' subject

Judgment sample- results when a researcher chooses the sample based on who they think would be appropriate for the study generally used when there is a limited number of people that have expertise in the area being researched

Multistage sample- combination of several sampling methods sampling frame| a list of individuals from which the sample is drawn

Under-coverage- occurs when some parts of the population are left out when the sample is selected

Bias- any systematic failure of a sample to represent its population which can include anything in a survey that may influence responses: wording of questions, interviewer's body language, etc.

Non-response bias- occurs when a large fraction of those sampled fail to respond

Observational study- a study based on data with NO manipulation of factors

Simulation- models random events by using random numbers to specify event outcomes

Statistic- value calculated for sample data - used to generalize to the population represented by the sample

Parameter- an attribute of the population that is estimated from the sample data

Math_OverviewBottomBanner.png IMAGES CREATED BY GAVS