Tips on implementing worksite smoking policy
surveys
- Overview
of Worksite Questionnaire: Content and Design
- Implementation Issues
- How
to calculate sample size
4. Other questions you have about conducting the worksite
survey
1.
Overview of worksite questionnaire: Content and
design.
The worksite survey covers the following 11 topics and has been tested by
our regional evaluation specialists.
1. Screening
questions - depending on the quality of your list, you can find the answers
to most of these prior to making the call
2. Number
of employees - important for analysis (size trends, size vs. policy –
for example larger companies may have a policy vs. smaller ones which
haven't formalized one).
3. Type
of industry, governance structure, who is responsible for setting smoking
policies - can help to guide your programming.
4. Are
employees and customers allowed to smoke? If so, where? (inside, outside,
vehicles) - nature of policies, level of smoking,
and where you can have an impact.
5. Does
the labor union agreement address smoking policies? - can be a powerful
leverage tool at larger sites.
6. Do
health insurance plans offer cessation benefits?
7. Has
worksite offered cessation assistance? - an area
that can be addressed with programming.
8. Is
there a formal vs. informal smoking policy
9. Reasons
for having/not having smoking policies - major motivations, can help target
campaign for changing attitudes - ex. cleanliness vs. health).
10. Employee
compliance with smoking policy and consequences if non-compliant - what's
the policy worth if no one follows it?
11. Assistance
to develop/improve policies. Information
you collect using these questions can be used to educate, inform, and to
help change policies.
- Implementation issues:
(These questions were raised by participants in our needs assessment
prior to the call)
- Refer to the
article entitled Implementing a
survey of worksite smoking policies: Fond du
Lac’s approach as a good starting
point. It is based on a conversation with Joyce Mann from the Fond
du Lac coalition
who has completed a worksite survey.
o
 Implementing
a survey of worksite smoking policies: Fond du Lac’s approach (2 pages, 444 KB)
Contacting the "right
person": The ideal person to contact for the worksite person is a
Human Resources manager. Look for someone who has some say in the policies
and some longevity in the company. If a human resources manager cannot be
reached, ask for someone knowledgeable about safety or smoking policies.
Also, a leadership office or the owner may know or may be able to direct
you to the correct person if they do not have the time to answer questions.
Although more time consuming, one can look at the company's website to
identify appropriate staff. This can also be a time to engage people by
offering to share results.
- Identifying
the "best" times to call worksites: Unfortunately there
is no "best" time to call a worksite. Basically, you have
to look at the county and the makeup of the work force to determine
what the best time would be. In general avoid the first thing Monday
morning and lunch hours. Also, call at night if the business is open.
If you call at the end or beginning of a business day, a VP or HR
Manager may have their phone ringing directly through to them.
- Experience
from those who have already implemented worksite surveys
Lynn Downing - Price County: Used initial design with 4 businesses.
- Small County:
Largest employer has 100 employees, with most at around 50 employees
- Most businesses
were quite happy to speak with the coalition.
- Answers were
fairly positive, yet guarded - It made them think.
- The employees
tended to follow (even unwritten) policies.
- No problems with
the length - they took from 5-10 minutes, the questionnaire looks
longer than it is.
- Worksites wanted
to make sure this information wasn't going to suddenly appear in the
newsapaer, etc. (To take care of this
comment - if you run into it, feel free to alter the introduction
sheet to suit your needs. For example, you might add something about
confidentiality, perhaps telling them that any data that is shared
will be stripped of identifiers).
- Q: How do we
coordinate doing the restaurant and worksite surveys simultaneously?
A: It really depends on your specific campaign. This is
something you might want to get individual help on, by contacting your regional specialist. In
general, the restaurant survey would be a better tool to focus on if
your campaign centered around exposure to ETS, but if you were dealing with employee
health, the worksite survey would be more helpful. Mailing one of the
surveys could be an alternative, but you run into validity and
response rate issues.
- Q: Is
terminating the survey for worksites with < 5 employees a
hard and fast rule?
A: No. You can adjust that number for what is appropriate for
your coalition. Don't use this cut off it if it doesn't apply. We are
seeking uniformity in our data (for a comparison between local level
data), but our primary goal is for coalitions to collect relevant
information for their campaigns. We will be able to sort things out
if you submit data on worksites with fewer than 5 employees.
- Q: If we use
questions from the survey, should we use your exact wording?
A: Yes, that would be a good idea. The surveys have been
carefully designed and tested to contain questions to help you in
your campaigns. If you do change the wording of questions, please let
us know before you submit data to us.
- Q: Are there
any questions we should not remove?
A: If you need to, you can remove questions that will not be
useful for you at the local level. Our caution is that although they
may not seem useful now, answers to these questions could be helpful
in years down the road. We tried to make the survey as short as we
could while still including questions that would assist you in
running your campaigns over the long term. If you need to cut
questions, please contact your regional
specialist for help.
- Check validity
of your responses: Whose response is credible? You may want to do
a small random sample and interview employees to see whether their
responses match those of management, particularly in worksites with
larger numbers of employees. May want to interview line workers or
ask for the mail room. This may show such differences as policy says
no smoking, but many of the people in the mailroom do smoke without
consequence.
How do we do this?
You will need to be creative to go about this. You could call the
operator and ask to speak to someone in one of these areas, and see
what happens. You may have trouble getting to the right people, but
many may be willing to speak to you. It may help to explain how the
information will be used, that it will be kept confidential, that
these are standardized unbiased questions, and that you are just
trying to see where they fall on reasons for and against smoking
policy. I would use one or two questions to do this - not the whole
questionnaire.
- Q: Should we
only target larger worksites, with 50 or more employees?
A: Kenosha's approach: Identify the top 10 largest employers
to see which have policies and get them to implement smoke-free
policies. This is a good approach to get information and fits their needs, however in general somewhere around 80% of Wisconsin
businesses employ < 50 people and these worksites (in total)
house a greater number of employees than a few large worksites (and
in turn can also affect a large number of people with ETS exposure). The approach is dictated by your
program goals, but in the long run it will be useful to collect as
much data as possible.
Rick Orton said Dane county did a survey of large worksites and got
good information. They found out most of the sites had a smoking policy, however this didn't provide them with any
information on the smaller employers. After reviewing their goals,
they found out that they really needed that information. He suggested
if you are already surveying larger employers, you should be
inclusive and try to get as much information as possible. It will help
you in the long run.
- Anticipate nonresponse: What about those who don’t
respond or refuse to participate? You need to choose a larger sample
in anticipation on nonresponse. Estimate
your telephone interview response rate at about 65%. To ensure
credible results, you should shoot for a 60% response in any survey.
That means if you want 122 surveys completed, you’ll need to
sample how many? 122 = 65% of what? 122/.65 = 187 in the sample. Your
sample size is 187 and your target response is 122, or 65%.
- Select your
sample: Using Excel, assign random numbers to each of the
worksites on your list. Generate a separate list at random. If your
list is smaller, cut up the names of the worksites, put them in a
hat, and draw out your sample of 187 or if you have a randomly
organized (non-alphabetized) list you can simply pick every 4th or
8th entry, or whatever interval will give you your sample size.
Types of Sampling to Ensure That you get the Information you
Desire:
A twist: Suppose that of that list of 1,427 worksites, only 5
worksites employ more than 300 people. These 5 worksites are the
biggest worksites in your county and together employ 4,700 people. If
you conduct a random sample of 187, you’ll
have 0.03% (one third of one percent) chance of selecting those
worksites in your sample. What can you do?
You can simply make sure you interview all the larger employers. This
is called oversampling. If you want
to make sure you have good representation from all worksites, you
will want to sort the worksites into groups according to size and
conduct a random sample of each group. This is called stratified
sampling, and it ensures that you will survey worksites of all
sizes. However, you will need to use the sample size chart for each
group, not for the whole population, which means you will have a
larger total number of interviews.
Adaption: Sampling can also
be adapted for use with mail surveys and focus groups. Depending on
the topic, length, etc., mail surveys may only have a 10%-40%
response rate and can introduce bias because of uncoded
responses (those famous responses that dangle out in the margin),
missing data, etc. This is the major reason the worksite survey was
designed to be a telephone survey tool. You will get a better
response rate and better data. Please contact
your regional evaluation specialist for help in converting the
worksite survey into a mail survey or if you have any sampling
questions or want to apply sampling to another tool.
5. Other questions you have about conducting
the worksite survey:
·
Q: How do we get our volunteers to reach
the "right" person at the "best" time?
A: The best way to help is to recognize the difficulties they will
be facing. Once they do 5 to 10 interviews, they will be more comfortable with
this. Maybe set up an incentive structure and reward the callers for
reaching a certain response goal. Also, be willing to complete
questionnaires when worksites call back.
·
Q: Should we exclude large worksites with
their parent companies out of state?
A: This is up to each individual program. It depends on your goals
and how much you think you can make a difference. This may be an issue for
border counties to discuss in more detail.
·
Q: Should we use Excel or Access to enter
our data? (which is preferred)
A: The Excel and Access instructions and templates for data entry
will be available on the web soon. In an August conference call we will
discuss analyzing the data more closely. Choosing either Excel or Access
will be up to each coalition, although we recommend that you use the Access
template to enter your data if possible as it allows more control over what
is entered. We hope the instructions and template we have created will be
easy enough to follow even if you are not yet familiar with Access.
|
