Carolina Chapter Job Bank: 1997 STC Salary Survey

by Michael Harvey, Data General

The following article is a three-part series published in the Carolina Chapter Communiqué.

Part I

More jobs seem available for technical writers than ever before. But just three years ago, the market for technical writers seemed sluggish. What, if anything, has changed during that time?

To try to get some answers, I polled the readership of the Carolina Communiqué in September of 1994 and also in January 1997. I reported the 1994 results, based on 67 responses, in a previous issue of the Communiqué. In a series of articles beginning with this one, I'll share the 1997 results.

Eighty-six Carolina Communiqué readers responded to the 1997 survey. As with the 1994 survey, I wanted to determine the salary range of our chapter's membership, as well as the age range, the gender breakdown, and the level of education. I also wanted to compile statistics about what we do, where we work, and how long we stay with a job.

For both surveys, I used the SAS system to statistically test the following questions:

• Does education predict salary?
• Does experience predict salary?
• Do contractors earn more than full-time regular employees?
• Do males earn more than females?
• Do older writers earn more than younger writers?

Of all the variables that I tested, only experience and education predicted salary in 1994. An analysis of variance confirmed that effect. In 1997, I performed similar analyses on the new data. This time, only experience predicted salary.

So what happened to eliminate the effect of education on salary? Did highly educated writers take lower paying jobs? Are there simply more technical writers in the area, with a wider range of education? In a future article, I will look more closely at the data and tell you what I find.

Still, the implication is the same now as it was in 1994: the more experience you have as a technical writer, the more money you make.

Take a look at this graph, which compares the salary distribution in 1997 to the distribution in 1994.



In 1994, the highest salary category I surveyed was "Above $50000/year." As you can see, that category had the biggest response. Wanting a more discriminating measure of this response, I added four salary categories above $50000/year in 1997, the highest being "Above $65000."

Notice that the 1997 and 1994 curves look similar, strongly suggesting two distinct groups of writers. Considering the results of my statistical analyses, I conclude that there is a less-experienced group making between $25000 and $50000 a year, and a more-experienced group making $50000 a year or more. I thought I would find where the salary range for the more well paid group peaks, but as you can see, I didn't. Notice too that the peak salary for the less-experienced group appears to have risen since 1994. I'll try to determine whether the increase is statistically significant.

Here's a graph that shows how many of us are contractors and how many of us are "permanent" employees.



Even though employment status had no significant effect on salary, it strongly correlated to job satisfaction in 1997. "Permanent" employees were more likely to report that technical writing met expectations in terms of job satisfaction than contractors were. Even so, all but four 1997 respondents said that they'd recommend technical writing to a friend. It's interesting that writers who are less than satisfied with the field would recommend it.

Finally, notice in this graph that roughly the same number of "permanent" employees responded both years, but that over twice as many contractors responded in 1997 than in 1994. Can this be a result of corporate downsizing?

Here's a graph showing the age distribution of the surveyed groups.



Look at how the tallest peak in the 1997 distribution is from 36-40 years of age, whereas in 1994, the tallest peak was from 31-35. The biggest disparity between 1997 and 1994 is for the 36-40 year group. Otherwise, the numbers in each group look similar.

The gender breakdown did not appear to change from 1994 to 1997, as the following graph shows.



Thirty-one males and 55 females responded in 1997. Twenty-six males and 47 females responded in 1994. The percentages are almost identical: 40/60.

Finally, the only thing that appears to be different between the education distributions is that there might be more folks with master's degrees in 1997.



In future articles, I will:

• Report what skills we think are important to our profession.
• Report where we work, how long we worked there, and where we worked before.
• Closely examine variables such as job satisfaction, professional growth, and tools used.
• More closely analyze the differences between the 1997 and 1994 data.

Detailed tables of descriptive statistics are also available.

Part II

Last month, I showed how experience predicted salary in my 1997 survey of 86 Carolina Communiqué readers. I compared my 1997 results to those I gathered in 1994. In both surveys, the more experience you had, the more money you made. This month, I'll begin by presenting the range of that experience.
Experience 1997 1994
0-5 years 19 23
>5 — 10 years 40 25
>10 — 15 years 13 9
>15 years 12 5
No response 2 5
86 67

The table shows that 1997 survey respondents are more experienced than 1994 respondents. In 1997, sixty percent more respondents have five to ten years' experience, 44% more have ten to fifteen years' experience, and 140% more have greater than fifteen years' experience. I don't think this is simply a function of our getting three years older, although that and an improving economy certainly would contribute to higher salaries. I think the numbers point to an influx of older and more experienced writers, and I believe that these more experienced writers brought higher salary expectations with them.

What skills do we think are more important to our profession: technical skills or writing skills? Hint: don't expect a simple answer from a technical writer.
Important skills? Frequency
Technical 9
Writing 34
Equal/both 26
Other 12
No response 5
86

Nearly 40% of us think that writing skills are more important, where only 10% of us think that technical skills are. But, being technical writers, and thus alert to the complexities of an issue, nearly a third of us thinks that both are important or are equally important. And it appears that 14% felt that the choices I offered were too constraining, and so reported that "desire" or "diplomacy" or "perseverance" were the most important skill to have.

What's the biggest change we've seen in the technical writing profession over the last five years?
Biggest change? Frequency
New technologies 15
Shift to on-line 12
Emphasis on technical skills 11
More contractors 5
More unqualified writers 4
Growth of field 4
Internet 4
Other 11
No response 23
89

Some respondents gave more than one answer, which accounts for the total exceeding 86. I distinguished between "shift to on-line," which comprises responses about more emphasis on on-line help than on printed manuals, and "Internet," which comprises explicit references to work on Web pages. "New technologies" comprises more general responses, such as the availability of more powerful desktop publishing tools.

The following table shows where we work, both now and in 1994.
1997 1994 Years there 1997 1994
Computer 28 28 0-5 years 55 47
Pharmaceutical 5 2 >5 — 10 years 18 12
Research 5 4 >10 — 15 years 4 3
Contract house 14 9 >15 — 20 years 2 3
Hospital 2 2 Over 20 years 3 1
Self 7 9 No response 4 1
Other 8 11 86 67
Unemployed 1 1
Telcom 14 0
Bank 0 1
No response 2 0
86 67

This table shows where we worked on our last job.
1997 1994 Years there 1997 1994
Computer 23 13 0-5 years 64 52
Pharmaceutical 6 1 >5 — 10 years 6 9
Research 9 6 >10 — 15 years 3 1
Contract house 19 24 >15 — 20 years 3 1
Hospital 3 1 Over 20 years 1 0
Self 3 2 No response 9 4
Other 17 17 86 67
Telcom 4 0
Bank 0 1
No response 2 2
86 67

We still work mostly at computer companies. More of us work for telecommunications and pharmaceutical companies now; it will be interesting to see whether and how this trend continues in a future survey.

The numbers suggest that in 1997, as in 1994, we don't stay at a job very long. Our reasons are varied, as the next table shows.
Why leave last job? Frequency
Better opportunity 22
Job ended 11
Relocation 10
Downsizing 10
Bad situation 7
Money 6
Career change 3
No response 5
Other 12
86


Here I made a distinction between "job ended," which included a contract terminating voluntarily, and "downsizing," which is never voluntary. The good news seems to be that many of us are good at managing our career, leaving a job for a better opportunity when it arises.

Finally, take a look at the tools that we use.
Tools Frequency
MS Word 39
Adobe FrameMaker 29
RoboHelp 10
CorelDraw 9
Adobe PageMaker 9
MS PowerPoint 8
Adobe Illustrator 7
Adobe Acrobat 6
Interleaf 6
HTML 6
BookMaster 5
MS Excel 5
WordPerfect 5
Adept Editor 5
Paint Shop Pro 5
Photoshop 4
Visio 4
HiJaak 4
Internal tools 4
Netscape 3
ASCII text editor 3
Adobe FrameBuilder 2
FreeHand 2
Adobe Distiller 2
MacDraw 2
SGML 2
No response 5
Other 19
210

The grand total exceeds 86 because respondents frequently listed more than one tool. I wasn't surprised that Microsoft and Adobe had such a strong hold on the desktop, but I was surprised that only two persons reported using SGML. I had been hearing so much about SGML over the last three years, I expected more readers to report it.

In my final article, I'll look at what affects job satisfaction, I'll examine whether the differences between the 1994 and 1997 results are significant, and I'll suggest avenues for further research.

Part III

In this, my final report on the 1997 salary survey, I will:

• Look at what affects job satisfaction
• Examine whether differences between 1994 salaries and 1997 salaries are significant
• Suggest topics for further research

Not surprisingly, writers report significantly more job satisfaction the more technical writing meets professional growth and salary expectations. Contractors report significantly less job satisfaction than "permanent" employees. This could due to job uncertainty, a perceived difference in status, or different types of work assignments. The topic bears closer scrutiny.

Interestingly, if writers produce new material or extensively rewrite material, they report significantly higher job satisfaction. I think this demonstrates that the more vigorously we exercise writing skills on the job, the happier we are with it.

According to my statistics, the following factors do not affect job satisfaction:

• Amount of training
• Resources available to do the job
• Extra hours worked
• Years of experience
• Highest degree obtained

The correlation between actual salary and job satisfaction was insignificant. Mind the distinction: the extent to which technical writing meets salary expectations correlates significantly with job satisfaction, but actual salary does not.

What does this suggest? Are our salary expectations low? Do technical writers prefer jobs that provide the opportunity to write new material, even when the job pays less than one that requires only light editing? Would writers take a lower paying job if it provided a better opportunity for professional growth? To answer these questions, we'll need more research.

The data I presented in my first article strongly suggested that writers are now making more money than they did in 1994. Does the upward shift in salary distribution reflect a real increase in earnings? The answer is yes. A chi-square test of differences between 1994 salaries and 1997 salaries was significant. I estimate that writers' salaries have increased about $3500 over the three year period. Should you schedule a heart-to-heart with your manager if your salary hasn't increased $3500 since 1994? I leave that to the reader to decide.

In addition to the research topics I raise earlier in this article, I recommend another complete salary survey, three years hence, to continue to track salary trends. Again, we must extend the upper range of salary categories to find the top of the second peak in the salary distribution. With local data from three surveys, we may be able to run interesting statistical comparisons to national data. We need to survey employers to determine what skills they think are important for technical writers, and compare results with those we gather from writers. And although I'm not sure how to do this, we should determine to what extent Carolina Communiqué readers represent the population of Triangle technical writers.

In short, we need to keep doing what all good technical writers do: keep asking questions.

---

Data owned by Michael Harvey. Please contact him regarding questions about the survey.