On, Off, or Asleep?: User Perceptions of Power Controls on Common Office Equipment

Azeen Chamarbagwala, Phoebe Kuan-Ting Liu, Craig Rixford,
Mary Trombley, and Bin Xin

SIMS, University of California, Berkeley, CA 94720

{azeenc, fibi, rixford, maryt, binxin @ sims.berkeley.edu}

This survey investigated existing consumer attitudes and perceptions concerning power control symbols and indicator lights on standard office equipment.  In addition to determining existing attitudes towards indicator light colors and states, sleep symbols, and power states, we also test users reaction to two different standards power state symbols: the existing international standard and one proposed by the Lawrence Berkeley National Laboratory (LBL).*  Results indicate that users found the proposed LBL standard more intuitive.  Users interpreted the proposed LBL standard in a more consistent manner than they did the current international standard.

Download PDF version of this document

I.  INTRODUCTION

Despite the widespread incorporation of energy saving features in office machines, such as copiers, monitors, and printers, the United States has failed to utilize this technology to its full potential. The Energy Analysis Department at the Lawrence Berkeley National Lab (LBL) estimates that $1.3 billion per year in energy costs could be saved in the United States alone if power-saving technologies were fully implemented (Kawamoto et al. 2000).  While energy saving features are installed in most office equipment, users often disable these features. 

Figure 1:  Estimated Enabling Rates of Common Office Equipment

Source: Nordman, Koomey, and Piette 2000.

Figure 1 shows the “enabling rates” of common office equipment.  Only 25 percent of all PC users fully utilize their computer’s energy saving features.  The LBL hypothesizes that these high levels of disabling are caused, in part, by the varied and confusing office equipment power control interfaces.  The LBL proposes that by making power management controls more intuitive and consistent across all office equipment, people will enable energy saving features more fully and thereby reduce overall energy consumption.  An international standard for the symbols does exist (IEC 60417 and ISO 7000); however, those few manufacturers that adhere to the standard often implement it in different ways.  In addition, there is an inconsistent implementation of indicator light color and states.  Figure 2 gives a few examples of the various symbols currently in use.

Figure 2: The Current Power Control Interface Standard

According to LBL researchers the U.S. could further reduce energy waste if a single, simple power management control interface was implemented on all office equipment.   To that end, the LBL group has developed a set of standards for these interfaces. The LBL has made the following five initial recommendations/propositions:

1. All machines have three principal power states:  on, off, and sleep.  Within any one of these states, devices exhibit consistent capability and behavior.  Furthermore, while a device may have more states, all other states can be classified or mapped into forms of the basic three states mentioned above.  (For example, the LBL group has tentatively identified the “Hibernate” state as a form of the Off state.)  This study accepts this proposition.
2. The word “power” should be used as standard terminology for the power(on/off) button. Once again, this study accepts this proposition.
3. The current international symbol for standby should replace the existing on/off symbol and a new international standard for sleep should be adopted.  This study will test this proposition.
4. Green/amber/off should be used to represent the three power states on, sleep and off respectively.  This study will test this proposition.
5. The sleep metaphor should be used for low-power states and the moon should be adopted as the new international symbol for that state. This study accepts the metaphor of sleep and tests the symbolic representation of it.

Thus for this study we assumed the first two LBL propositions as well as the metaphor of sleep to describe low-power states. To test the LBL group’s hypothesis, we conducted a survey that gauged existing consumer attitudes toward and understanding of power controls on common office machines - assuming the first two LBL propositions.  Our survey also tested respondents’ comprehension of indicator light colors and states (solid or blinking).

Specifically, we wanted to answer the following questions:

To test whether the current international symbol for on/off should be replaced by the existing standby symbol and a new international symbol for sleep be adopted.

• Do users understand the current international power control standard?
• Would users more easily understand the new, LBL-proposed, standard?
• Is the current international symbol for standby easily interpreted by users across all devices?

To test whether green/amber/off should be used to represent the three power states on, sleep and off respectively.

• How do users ascertain the power state of an office machine?
• Are there consistent user expectations (across devices) for indicator light color and state to specify the power state of a device?

To test whether  the moon should be adopted as the new international symbol for that state.

• In general, what are users’ current perceptions of power saving features?
• Do users understand the current international symbol for sleep/stand by?
• Is there a more readily-understood symbol to indicate the sleep state?
• How do user’s interpret the moon symbol?
  

II.  METHODS

A.  Participants

In this study we surveyed 37 non-colorblind UC Berkeley students, over the course of two weeks, in October 2001 using an online, multiple-choice survey.  We chose UC Berkeley students because they were readily available and could be assumed to have some familiarity with office equipment.  Users were given a $10 gift certificate after the survey was completed.

B.  Survey Implementation

We first conducted a pilot survey with 10 respondents.  The pilot survey asked open-ended questions about the respondents’ experience with using power controls and elicited information about the lights and symbols they associate with various power states.   Their responses helped guide the design, selection, and wording of questions and multiple-choice responses in our survey.  For the final survey, we utilized an online form to facilitate data collection and to enhance our ability to display appropriate graphics.

Our survey questions can be divided into three sections. The first asks general questions about the respondent’s experience with energy saving features (questions 1-7).  The second section asks about existing perceptions of or associations with power states, symbols, and lights (questions 8-19).  The last, more experimental section attempts to gauge how a user would expect the machines to change from one power state to another (questions 22-27).  See Appendix 1 for links to each of the surveys.

Table 1 : Outline of Survey Questions

C.  Study Design

We utilized both a within-subjects and between-subjects design.  All respondents answered the General Attitudes questions at the beginning of the survey.  We did not vary the order of the questions within the survey.

Effect of Device: In order to assess whether responses were dependent on the context of a specific device we chose two different types of office equipment for questions 8-22 in our survey: a laptop computer and a copy machine.  Half the subjects were given a set of questions that referred to a laptop (group A); the other half received questions that referred to a copy machine (group B).

Effect of Standard: In addition we compared two different sets of symbols-- the existing international standard and the standard proposed by the LBL group.  All subjects were asked questions using both symbol sets; however, we varied the symbol set they viewed for any given question. 

Figure 3 : Current international standard and proposed LBL standard

            This survey design thus required us to administer four different versions of the final survey.  (See Appendix 1 for links to the online surveys.)

Table2 : Final Survey Design

III.  RESULTS

Survey results show that users interpret and act in a more consistent manner when presented with the proposed LBL standard than when presented with the existing international standard.  To the extent that the behavior of the power controls can be implemented to take advantage of this consistent user behavior, our study supports the LBL’s contention that a revised power controls standard will increase the ease of using energy-saving features.  Although users did not strongly associate the moon symbol with sleep, when it was presented in juxtaposition with LBL’s proposed power symbol, they were very likely to associate the symbol with the sleep state.

A. General Attitudes

Disabling Energy-Save Features

Users were asked whether they had ever used energy saving features, and whether they had ever disabled these features and, if they had, why they did so.

Out of the respondents that had used energy saving features, a large proportion, 38 percent, reported disabling these features (see figure 4).  Interestingly 32% report not having used energy saving features.  Most likely they have used machines with these features but have not altered their configuration: all energy star labeled equipment is supposed to be shipped with these features enabled.

Figure 4:  Percent of respondents that have used / turned off energy saving features

Respondents’ reasons for disabling energy saving features varied.  The top three answers were:

(1) The device goes into power-save mode too soon (30%)
(2) It takes too long for device to come back on (22%)
(3) Use results in system errors or crashes (22%).

Figure 5:  Reasons for turning off energy saving features

Almost half of the responses can be attributed to the failure of users to understand or correctly set the energy saving features and functions (device goes into powersave mode to soon and unsure of how to use features).  These data seem consistent with the LBL’s contention that improving power controls interfaces would likely increase their use.

B. Replacing Current International Standard Symbols

Expected Behavior of Machines: Button only manipulation of device state

One survey section focused on examining how a user expected a device to respond to pushing a power state control button.  Respondents were asked to move a device (laptop or copier) from a given state to another state, using only button symbols.   No captions or illustrations were provided, with the exception of the button symbols. Each person was asked to change the state of the device using buttons with text labels and then with labels having one of two symbol sets: the current international standard or the LBL’s proposed standard.

We asked half the respondents to answer the questions in the context of a copy machine and the other half for a laptop.  We conducted a two-way contingency table analysis to evaluate whether people answered the question differently depending on the device type.  Once again we found their answers do not appear to be dependent upon device type:

• To change from ON to SLEEP: Pearson X² (2, N=36) = 3.91, p=.141
• To change from SLEEP to ON:  Pearson X² (2, N=37) = 2.39, p=.302
• For a machine in the SLEEP to OFF state: Pearson X;² (2, N=36) = 4.19, p=.123

We assume, therefore, that there are not significant differences between the devices and analyzed the responses about transitions for both devices together.  See Appendix 2 for details of our chi-square analysis.  The tables below show our crosstabulation tables and chi-square results for questions 20-22. The figures below (figures 6-8) summarize which buttons respondents believe will trigger the desired change of state in the device. 

Turning the device from On to Sleep:

 LBL symbols elicited a consistent response from users 79% of the time. International symbols were less intuitive to users, 55% of respondents were unable to identify which button to press to change the state of the device.

Figure 6:  Which Button Would Turn the Device from On to Sleep

Turning the device from Sleep to Off :

LBL symbols elicited a consistent answer from users 55% of the time.  International symbols led users to give the correct answers only 39% of the time.

Figure 7:  Which Button Would Turn the Device from Sleep to Off

 Turning the device from Sleep to On:

In this case users selected a consistent response from users 62% of the time compared with a 56% response for the International standard.  In this case, however they chose the same symbol: the proposed LBL power symbol.  Although the consistency of the responses is about the same under both standards, it is the “incorrect” choice

Figure 8:  Which Button Would Turn the Device from On Sleep to On

Respondents’ perception of device power state and expectations of how a device will change states through power control buttons

The final questions in the survey were illustrated with laptops depicting devices in various power states.  Users were not informed of the state of the device but had to deduce it from the screen content (dark or active) and the indicator lights (of different colors and states).  The survey question, as with the section above, asked users to put the laptop into a desired state (e.g., turn it on, turn it off, put it to sleep). 

First, we asked respondents to choose what they would do to put the laptop into the desired state.  We offered them a list of typical actions such as: “touch the space bar” or “use software controls.”  The following table summarizes how users would typically accomplish the given task.

Table 3:  What Action Users Would Take to Bring About Change of Power State

Next, respondents were asked what they would do if they could only alter the laptop’s state by pushing one of the two power control buttons.  Half the respondents for each question were presented with the existing international standard while the other half were presented with the LBL’s proposed standard. Table 4 shows the percent responses for each button to the given question.  (Note that for those questions that ask the user to put the device into an identical state, the existing state is depicted differently.)

Table 4: Which Button Users Would Use to Move Laptop into Desired State

The following figure shows examples of the images a user might see in these questions (at a reduced scale). 

Figure 9: Examples of laptop images that indicate state of the laptop

Figure 10: Which Button Would You Push to Turn the Laptop On?

State of device: dark screen with solid yellow light. 

Objective: turn machine on

Results: Figure 10 shows that given a laptop device that appears to be in some low power state over 50 percent of respondents in both groups select the same symbol.  In the case of the international standard their choice is technically incorrect (although in reality it may be the correct choice since manufacturers do not necessarily adhere to the standard).  In addition, about 40 percent of users faced with the international standard did not know which button to choose, nearly double that of the proposed LBL standard.  This suggests that users have a clearer sense of the difference between the buttons in the LBL standard.

Figure  11: Which Button Would You Push to Turn the Laptop On?

State of device: dark screen with blinking green light. 

Objective: turn machine on

Results: The responses to this question were similar to responses to question previous question.

Figure 12: Which Button Would You Put the Laptop to Sleep?

State of device: active screen with solid green light. 

Objective: put the laptop to sleep

Results: Close to 80 percent of those looking at the LBL standard chose the moon symbol to put the laptop machine into the sleep state.  Of those looking at the current international standard, 35 percent chose the on/off symbol and 50 percent did not know which button to choose.

Figure 13: Which Button Would You Push to Turn the Laptop Off?

State of device: active screen with solid green light. 

Objective: turn the laptop off

Results:In this question, over 90 percent of those using the LBL standard chose the power symbol compared with 60 percent for this existing standard’s on/off symbol.  That 60 percent chose the correct international symbol is inconsistent with other results in this survey (which suggest people are familiar with the standby symbol being used for power).  This result could, perhaps, be explained by the fact that people generally use software controls to turn computers off.

Figure 14: Which Button Would You Push to Turn the Laptop On?

State of device: dark screen with no light. 

Objective: turn the laptop on

Results: In this question, over 90 percent of those looking at the LBL standard chose the power symbol compared to the 18 percent who chose the international on/off symbol.  Also, the number of people who don’t know which button to push suggests that the juxtaposition of international symbols is not as intuitive as the proposed LBL standard.

C. Representing Power States with Colors

Our survey included questions to collect information on the means by which people ascertain the power state of a machine.  Figures 15 and 16 illustrate the cues and actions by which people determine the existing power state of a laptop computer or copy machine.

Figure 15: Cues and Actions Used to Identify the Power State of a Laptop

Figure 16: Cues and Actions Used to Identify the Power State of a Copy Machine

The results indicate that neither machine’s power control indicators provide an obvious way to assess its state.  Users react to this fact by using a variety of cues and actions to determine each device’s power state.  The copy machine seems to be more straightforward to assess with 74% of respondents stating that they would look at either indicator lights or the control panel to assess the machine’s state.

In contrast, 16% of respondents would move a mouse and 59% of respondents would press buttons to determine the state of a laptop.  In the case of the laptop, users perform an action and then await the machine’s response to infer the state of the machine.  With the copier, users look directly at indicators of the machine’s state. We can infer that no one method of understanding the state of a machine is prevalent in our population, but that, perhaps, copy machines have more easily discernible power state indicators.

Existing associations between power states and indicator lights

Our survey also examined respondents’ existing attitudes towards indicator lights and machine power states.  As mentioned in the methods section above, we asked half of the respondents questions in the context of a copy machine and the other half in the context of a laptop computer.  For questions 8-10, we conducted a two-way contingency table analysis to evaluate whether the color a user associates with a power state depends on whether the device was a laptop or copy machine.  The choice of color associated with the power state was not found to be significantly related to machine type:

• For a machine in the ON state:  Pearson X² (3, N=35) = 3.62, p=.306. 
• For a machine in the OFF state:  Pearson X² (4, N=35) = 5.44, p=.245
• For a machine in the SLEEP state:  Pearson X² (4, N=33) = 5.11, p=.276

Therefore, for the purposes of this study we assumed that there weren’t significant differences between the devices and analyzed the responses about color association for both devices together.  (See Appendix 2 for the details of the chi-square analysis.)  The figures below illustrate the users perceptions of color and power state. 

Figure 17: Color Association for an On Button

Figure 18: Color Association for an Off Button

Figure 19: Color Association for a Sleep Button

These results show that while users have fairly consistent color associations for on and off, no clear association exists between the sleep state and a color.  In fact, the most popular choice, blue, is rarely used as an indicator light on office machines.  While people may associate the color blue with the sleep state, it is likely that this association does not come from their experience with office equipment. 

Indicator light and state association

A set of questions in the survey asked what power-states the user associated with a particular color light (both solid and blinking) in order to determine whether respondents already had very strong associations with particular indicator lights.  (The blinking lights were depicted by animated gifs of same color as solid indicator that cycled at .5 seconds on .5 seconds off.) The following figure summarizes the data gathered in this section of the survey.

Figure 20:  Status of Machine (including power states) Associated with Various Colors

With a few exceptions, our respondents did tend to associate particular power states with the non-flashing indicator lights.  For example, the overwhelming majority of users believed that a solid green indicator light meant that a machine was on.  While 50% of respondents believed that an orange light indicated the sleep state, 32% of respondents believed that yellow indicated the sleep state.  Combining the two categories of responses into a yellow/orange category, we find a high level of yellow/orange association with sleep.  Our survey results indicate that a large percentage of users-- about sixty percent-- do not associate a solid read light with any particular meaning. These results suggest that the LBL’s proposed standard of green, amber, and no light to represent on, sleep, and off would be readily understandable to existing users of machines.

Users did not have strong associations with blinking lights of most colors.  However, nearly 50 % of all users believed that a blinking red light meant that a device was in an error state.  These data met our expectations, as red is a warning color for many common devices. When respondents had an opinion on the other blinking colors, they seemed to believe that blinking indicated some kind of special state-- that the machine needed input, was in transition, or was in an error state.

D. Choosing a Symbol for the Sleep State

Respondents were asked to rank how strongly they associated a number of symbols with the sleep state in order to understand how the LBL’s proposed symbol - the moon - would compare to other possible symbols.  The set of possible symbols we presented in our survey were culled from the responses to our preliminary survey.

The following figure gives the average result for each symbol (responses were given as a ranked scale from 1-5).

Figure 21: Strength of Association of Sleep State with Various Symbols

The text button (unsurprisingly for English speakers) has a very strong association with the sleep state.  That the word sleep has only an association of 4.3 out of 5 suggests that the moon’s rating of about 3 should be considered more positive than neutral.  As noted previously, the symbol of the moon juxtaposed with the LBL power symbol, and presented in the context of power controls, seems to effective.  Nevertheless, we would suggest more exploration of potential symbols and testing to ensure that it works across cultures. The results indicate that the proposed moon symbol has a neutral association with the sleep state. The symbol of a person in bed—already in use in travel guidebooks as a symbol—has a strong association with sleep. 

IV.  CONCLUSION

Our survey results indicate that many users (38%) have disabled energy saving features on office equipment.  These results support the LBL’s hypothesis that significant amounts of energy are wasted due to these features being disabled on office equipment. Changes to the power controls interface might help to remedy this problem. 

Our data suggest that users do not respond to different devices in different ways.  That is, a user will not respond differently to a power management control interface based on the type of machine to which it is attached.  This is not to say that users don’t interact with the devices differently, rather that they don’t interpret buttons or colors to mean different things for different machines. This finding supports the LBL’s argument for a single standard for all office equipment.

The data for existing associations with colors are varied.  Overwhelmingly, green is the choice for the indicator light for a machine in the ON state, independent of machine type.  The highest choice for the OFF state was “No Light” for copy machines, but laptop respondents chose “No Light” and red in nearly equal measure.  Since our respondents had a wide range of opinion regarding the color for the SLEEP state, that we recommend more refined testing for this category.

In general, respondents had stronger state associations with solid lights.  With the exception of the blinking red light, users did not really understand what blinking lights meant.  This is in marked contrast to the solid lights, where associations for green (ON) and yellow/orange (SLEEP) were fairly consistent.  Because users’ interpretations of blinking lights are mostly ambiguous, blinking lights should be incorporated into power control standards with care.

Results indicate that users do not strongly associate the moon symbol with the SLEEP state (in contrast to text symbol or even the man in bed with ZZZZs).  However, this does not imply that the moon symbol should not be used.  It simply suggests that further consideration and investigation of the most appropriate symbol is necessary.

Overall, the LBL’s proposed button symbols proved more intuitive to users.  Respondents’ high percentage of consensus with the LBL’s standard across device type suggests that respondents are better able to comprehend the LBL’s button symbols.

V. REFERENCES

Kawamoto, Kaoru, Jonathan Koomey, Bruce Nordman, Richard E. Brown, Mary Ann Piette, and Alan Meier. 2000. “Electricity Used by Office Equipment and Network Equipment in the U.S”. Published in the Proceedings of the 2000 ACEEE Summer Study on Energy Efficiency in Buildings. Asilomar, CA. August.

Nordman, B., A. Meier, M.A. Piette. 2000. "PC and Monitor Night Status: Power Management Enabling and Manual Turn-off." In Proceedings of the 2000 ACEEE Summer Study on Energy Efficiency in Buildings, 7:89-99. Washington, D.C.: American Council for an Energy-Efficient Economy.

APPENDIX 1:  ONLINE SURVEYS

http://fusion.sims.berkeley.edu/power_save_survey/SurveyA.htm

http://fusion.sims.berkeley.edu/power_save_survey/SurveyB.htm

http://fusion.sims.berkeley.edu/power_save_survey/SurveyC.htm

http://fusion.sims.berkeley.edu/power_save_survey/SurveyD.htm

APPENDIX 2:  CHI-SQUARE ANALYSIS

To explore whether or not users have different associations and expectations for power controls when using different equipment, we had half of our respondents answer questions in the context of a copy machine and the other half in the context of a laptop computer.  Prior to analyzing the data we performed a chi-square text to determine  whether the responses of the two groups differed significantly.  

Color Association with three power states

Are people’s associations of the color of indicator lights to three different power states (on, off, sleep) machine specific?  We conducted a two-way contingency table analysis to evaluate whether people associated different colors with the same state depending upon whether the machine was a laptop or copy machine.  The choice of color associated with state was not found to be significantly related to machine type:

• For a machine in the on state, Pearson X² (3, N=35) = 3.62, p=.306. 
• For a machine in the OFF state, Pearson X² (4, N=35) = 5.44, p=.245
• For a machine in the SLEEP state: Pearson X² (4, N=3S3) = 5.11, p=.276

Therefore, for the purposes of this study we assumed that there weren’t significant differences between the devices and analyzed the responses about color association for both devices together.  The tables below show our crosstabulation tables and chi-square results for questions 8-10

Question 8: What color do you associate with a copy machine / laptop being on?

Crosstabulation

Chi-Square Test

a.  6 cells (75.0%) have expected count less than 5. The minimum expected count is .49.

Question 9:  What color do you associate with a copy machine / laptop being off?

Crosstabulation

Chi-Square Tests

Question 10:  What color do you associate with a copy machine/ laptop being in the sleep state?

Crosstabulation

Chi-Square Tests

Association of button with transition between states:

For another part of our study we asked questions about which button would move the device from one power state to another.  We asked half the respondents to answer the questions in the context of a copy machine and the other half for a laptop.  We conducted a two-way contingency table analysis to evaluate whether people answered the question differently depending on the device type.  Once again we found their answers do not appear to be dependent upon device type:

• To change from ON to SLEEP: Pearson Χ² (2, N=36) = 3.91, p=.141

• To change from SLEEP to ON:  Pearson Χ² (2, N=37) = 2.39, p=.302
• For a machine in the SLEEP to OFF state: Pearson Χ² (2, N=36) = 4.19, p=.123

We assume, therefore, that there are not significant differences between the devices and analyzed the responses about color association for both devices together.  The tables below show our crosstabulation tables and chi-square results for questions 20-22.

Question 20:  Which button would you push to put the copy machine/laptop into the sleep mode [from the on state]?

Crosstabulation

Chi-Square Tests

Question 21:  Which button would you push to put the copy machine/laptop into the on state [from the sleep state ]?

Crosstabulation

Chi-Square Tests

Question 22:  Which button would you push to put the copy machine/laptop into the off state [from the sleep state]?

Crosstabulation

Chi-Square Tests