Dementia is a major public health concern but one that continues to be stigmatised. We examine lay knowledge of dementia and attitudes to people with dementia as potential precursors of public anxiety, focusing on the social characteristics associated with (a) the formation of these attitudes, and (b) the perception of the need for restriction and control for people with dementia.
Analysis of the 2014 Northern Ireland Life and Times survey, which included questions on knowledge of, attitudes to and personal experience with dementia. We used (a) latent class analysis and (b) logistic regression to examine factors associated with respondent attitudes towards dementia.
Respondents (n = 1211) had relatively good general knowledge of dementia, but limited knowledge of specific risk factors. Negative perceptions of dementia were mitigated somewhat by personal contact. A high proportion of respondents felt that high levels of control were appropriate for people diagnosed with dementia, even at early stages of the disease.
Personal antipathy to dementia was highly prevalent despite ongoing public campaigns to increase public awareness of developments in its prevention, treatment and consequent care pathways and hampering efforts to widen social inclusion. Fresh thinking and more resources may be needed to challenge persisting common misapprehension of the condition and the formation of entrenched stigma.
Citation: Rosato M, Leavey G, Cooper J, De Cock P, Devine P (2019) Factors associated with public knowledge of and attitudes to dementia: A cross-sectional study. PLoS ONE 14(2): e0210543. https://doi.org/10.1371/journal.pone.0210543
Editor: Stefano Federici, Università degli Studi di Perugia, ITALY
Received: October 7, 2018; Accepted: December 27, 2018; Published: February 28, 2019
Copyright: © 2019 Rosato et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying the results presented in the study are available from the Northern Ireland Life and Times Survey (NILT). Details of these data base, including how to access the data, are available by accessing the website: http://www.ark.ac.uk/nilt/.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Although approximately 30% of people who reach the age of eighty are likely to have some form of dementia  evidence suggests that public knowledge of the disease is limited [2–4]. While initiatives such as dementia friendly health-care settings and communities [5,6] seek greater social inclusion for people with dementia there is scant evidence of their success. The World Health Organization has called for national public health agendas to highlight dementia as a priority, especially in relation to prevention, early detection and intervention [7–9], and including calls for greater public awareness of dementia in general, and increasing lay knowledge of modifiable risk factors [10,9]: improvement in both are key to reducing dementia-related stigma.
Current research in dementia-related stigma reveals social isolation  and discrimination, even within medical care settings . It may partially explain deficits in rates of early dementia diagnosis in the United Kingdom  acting as a barrier to help-seeking , emotional and psychological distress and social exclusion. Werner et al  suggest that stigma felt by family caregivers increased their negative experience of caregiving.
Stigma is the assigning of a socially discrediting stereotype by the wider society which provokes an individual to feel rejected in some way—a spoiled identity . The concept has a long history in social research and has been applied to analysis of, for example, ethnicity, sexuality and health-related contexts such as human immunodeficiency virus (HIV/AIDS) and disability . In recent years there has been increasing public awareness about dementia  including a trend for more in-depth sympathetic representations . However, anxiety about dementia-related agitative behaviour may result in persisting high levels of social exclusion [20, 21, 22]. Additionally, an emphasis in institutional care settings on containment and control, rather than person-centred approaches, can be seen as reasonable and therefore accep and more likely. Surveillance or assistive technology is promoted as a means of monitoring people who exhibit agitative behaviour, including wandering–a distressing, poorly understood problem often the reason for nursing home admission [23,24]: a solution maybe, but one with significant ethical and human rights implications . This study examines both current public knowledge of dementia and attitudes to people with dementia using the Northern Ireland Life and Times Survey . In particular, we examine the socio-demographic characteristics associated with (a) lay knowledge of dementia, (b) lay assumptions about dementia, and (c) attitudes to control and coercion thought appropriate for people with dementia.
Materials and methods
Study setting, design and participants
The Northern Ireland Life and Times (NILT) survey is a cross-sectional attitudinal survey, undertaken annually in Northern Ireland (NI) since 1998, involving a random sample of 1211 adults aged eighteen years or more. Full details of the survey design and methodology are outlined in the published Technical Notes . Briefly however, a two-stage sampling process was used, as a comprehensive sampling frame of individuals was not available to the researchers: a systematic random sample of 2,449 addresses was selected from the Northern Ireland Postcode Address File (with business addresses excluded); and, using the next birthday rule, the interviewee was randomly selected from those aged eighteen or over living at the address. The main fieldwork was carried out from 22/09/2014 to 26/12/2014. More general notes on the survey are available elsewhere [27,2]. Two questionnaires were devised for use in the 2014 survey—both using an Ipad for data entry: a main face-to-face interview; and a shorter self-completion questionnaire for more sensitive issues (although the interviewer could record responses if requested). Both are accessible from the main NILT website (www.ark.ac.uk/nilt), and this analysis is based on the former. Because the question modules change from year-to-year ethical approval is obtained annually: the 2014 survey received ethical approval from the Ethics Committee in the School of Sociology, Social Policy and Social Work, Queen’s University Belfast (http://www.ark.ac.uk/nilt/2014/).
Data comprised a number of questions assigning socio-demographic attributes–including: age (grouped for this analysis as 18–34, 35–54, 55–74 and 75 years or older); gender; highest educational qualification (degree-level, intermediate level, and no stated qualifications); occupational social class (professional, intermediate, routine social classes, and other–i.e. persons classified to a social class by means other than occupation); economic activity (summarised as active, inactive); personal and household income (separately, each grouped into four bands from highest to lowest income levels); marital status (currently married/cohabiting, never married, with those widowed, separated or divorced combined as a single category); housing tenure (owner occupier, private renter and those in social rented housing); and finally, the locale in which the respondent lived (urban, larger towns and rural areas). Also asked were questions on residential attributes: how long the respondent lived in their area of residence (less than 10 years; 10–19, 20–39, and 40 or more years); and whether they had ever lived outside NI for more than six months (never outside NI; lived in Great Britain or the Republic of Ireland, or lived elsewhere). Several questions were included on religion and religious observance: religion stated (yes, no); religious denomination (Catholic, Protestant, Other); and church attendance (at least weekly, monthly, yearly, or never attend).
Ten questions related to personal experience and knowledge of dementia, personal attitudes to dementia and perceived public perceptions. These are detailed below [28,2]. Two questions examined the form and intensity of respondent contact with people with dementia: (a) on level of contact—none, job-related, within family or as part of some wider circle of acquaintance; and (b) for those recording such contact, on the level of caring experience or assistance they might have given. For analysis purposes these were recoded as: contact (none, contact within family/friends, or amongst acquaintances); and experience of caring (none, direct experience of caring, or offering other assistance).
Knowledge of dementia was examined across two dimensions: as (a) more general or tacit background knowledge (is it a disease of the brain, a mental illness, part of normal ageing, or another term for Alzheimer’s), and (b) more specifically, associated risk factors (is there a genetic component, is it related to heavy smoking, poor diet, high blood pressure or excessive alcohol consumption). The responses to each set of questions were aggregated into three categories (agreement with the statement, disagreement, or no opinion offered) and each separately further summarised: for (a) as the number of correct answers to the questions; and for (b) as four or more incorrect, three incorrect, a mixed response (for the five answers in the ratio 2-2-1), three correct and four or more correct.
Negative and Positive attributions of dementia: respondents were provided with thirteen words relating to dementia (unpredictability, fun, gentility, fear, confusion, lost, kind, dangerous, trapped, happy, angry, sad, pathetic or other—each allowing only yes/no (0/1) responses) and asked to select those most representative of the way someone who has had dementia for a long time appears to them. Using this we derived a typology of attitudes to dementia based on the patterns found in the word associations, and represented in terms of levels of positivity or negativity towards the state of dementia.
Perceptions of Dementia recorded the level of respondent agreement with statements on their perceptions of dementia, in terms of both personal autonomy and treatment of those with dementia, and depletion of their character or personhood.
Perceptions of control or personal autonomy: a measure of how much personal autonomy respondents thought it appropriate to afford people with either early-stage or late-stage dementia. This was based on four questions—should they continue to live unaided, control their own medication, continue to drive a car or wear an electronic tag: with responses summarised as agree, disagree and no opinion offered. All except the information on electronic tags were combined and a composite variable derived quantifying the extent to which respondents thought it was appropriate to curtail or control the personal autonomy of those with dementia: with values (a) low levels of control appropriate; (b) no opinion offered; and (c) high levels of control necessary.
Caregiving in dementia: examined respondent opinion on the effects of caring for someone with dementia–whether it can be lonely, rewarding; or can affect caregiver’s health. This was summarised as a single variable: respondents agree, disagree or record no opinion on what they perceive to be the effect of caring for those with dementia.
We described the main socio-demographic features of the respondent cohort, their personal experiences, perceptions and background knowledge of dementia and examined these in relation to the main outcomes (outlined above) using multinomial logistic regression. Using the 13-word item question, we undertook latent class analysis (LCA), deriving a summary three term classification of respondent perceptions of dementia, as captured by the word associations–here the LCA technique groups the participants on the basis of similar response patterns in their selection of the word associations, and generates new indicator variables based on these. The selection of an optimal latent class solution (or Best Fit) is evaluated using a standard set of fit statistics: the standard Akaike Information Criterion (AIC) and the Bayesian Information Criteria and its sample-size adjusted version (BIC & SaBIC), as well as the Mendell-Rubin Likelihood Ratio Test; with the model selection also assessed in line with both substantive rationale and background theory). Finally, logistic regression was used to examine factors associated with attitudes to dementia. The main analysis was undertaken using Stata 13 , with Mplus  used for the LCA.
Of 2,161 potentially eligible respondents 481 could not be located and 469 declined involvement, leaving 1,211 respondents, a response rate of 56%, similar to earlier studies . Table 1 shows selected socio-demographic and socio-economic characteristics for the respondent group, with column 5 comprising a set of tabulations derived from the Northern Ireland Longitudinal Study  (and see note at the foot of Table 1). While these while are presented as indicative only the majority of the indicators compare well to that of the study population: two exceptions being the older age structure of the respondent population when compared to the Northern Ireland population; and the distribution of housing tenure with fewer of the respondent population in owner occupation and correspondingly more in social rented accommodation.
Contact and care-giving
Females and people aged 35–74 were more likely to have contact with people with dementia irrespective of locale, with owner occupiers more likely to report within-family contact (74%) than other tenure groups, and Catholics less likely than Protestants to report within-family contact (43% and 52% respectively). Of those reporting within-family contact, 39% had a caring role and 43% a lesser helping role. Among people with more distant contact 33% reported no experience of caring and 53% provided help other than caring. Generally, while 64% of respondents showed relatively good lay knowledge of dementia, 78% incorrectly answered the more specific risk-factor questions (Table 1). A substantial proportion of people cited having no opinion (ranging from 39% to 56% for each question), and 56% of respondents wrongly answered four (of five) questions asked. Only 11% answered four or more correctly (Table 2).
Table 3 records perceptions of the experience of caring for people with dementia and its potential effects on the caregiver. Caregiving was regarded as predominantly burdensome (67%): it can be lonely (80%), unrewarding (52%) and potentially a health burden (71%). Table 4 examines respondent perceptions of how dementia affects those diagnosed—with the majority either expressing no opinion or espousing statements highlighting aspects of loss-of-self (44.6% and 49.4% respectively).
Attributes of dementia
Relationships between thirteen putative dementia attributes elicited more complicated responses: eighty respondents didn’t answer and thirteen included all attributes—these were excluded, leaving 1118 for analysis. The most commonly chosen attribute was confused, cited by over 90% of respondents and also most consistently selected in conjunction with other items. Generally, more positive or empathetic attributes (fun, gentility, kind and happy) correlated highly with each other but lowly with attributes classifiable as more negative. The term pathetic was cited least in conjunction with the other attributes (ranging from 17% to 30% over each) and the term dangerous was cited most often in conjunction with attributes eliciting empathy (range = 36%-46%). We used LCA to establish relationships between these attributes—Fig 1 shows the emergent three-class solution, indicating the feelings involved in thinking about the lived-situation of people with dementia: a small group reporting mixed positive/negative feelings (4% of respondents (n = 50)); a group recording high levels of negativity (26%); and a larger group recording more moderate levels of negativity (70%). The groups exhibiting moderate attitudes were combined producing a binary outcome: those with moderate attitudes, and those describing high levels of negativity (0/1). Table 5 shows the relationship between these expressed negative attitudes, contact with people with dementia and knowledge of dementia. In the fully adjusted model: females more likely than males to express negative attitudes (Odds Ratio (OR) = 1.71: 95%CI = 1.28, 2.29); similarly, those in urban areas compared to their rural counterparts (OR = 1.42: 1.05, 1.94); while no association was recorded with age (however, those in the eldest age group recording lower likelihoods when compared with those younger). Respondents with higher levels of general knowledge of dementia and those less knowledgeable about risk factors were more likely to express higher levels of negativity (OR = 2.51: 1.64, 3.83 and OR = 1.71: 1.19, 2.46 respectively) than their comparator groups, as were those who had personal contact, compared with no contact (OR = 1.59: 1.01, 2.50 for when within a wider circle of acquaintanceship).
Autonomy and control
Eighty per cent of respondents thought it appropriate that high levels of control be exerted on those with late-stage dementia, with 43% regarding it as appropriate even at early-stage (Table 6). Those offering no opinion to these questions also followed a pattern–higher proportions offered no opinion to early-stage than late-stage scenarios (39% and 18% respectively). Most respondents were disinclined to allow people with dementia to drive (59% and 84% for early-stage and late-stage respectively), and most considered electronic tags necessary in both early- and late-stage scenarios (53% and 57% respectively).
Rural residence, lower education and age were associated with toleration of tagging, even in early stage dementia (Table 7). Conversely, those with personal contact and experience of caring, and knowledge of dementia were less likely to advocate tagging. Columns 5&6 highlight factors associated with acceptance of the need for greater control of people with early-stage dementia. Early middle-aged respondents (35–54) were less likely to advocate control (OR = 0.55: 95%CI = 0.35, 0.87), and those less well-educated more likely, compared to people with degree-level qualifications (OR = 1.56: 1.07, 2.28 and OR = 2.98: 1.76, 5.03 for intermediate-level and no qualifications respectively); and those with more limited knowledge of risk factors (OR = 2.75: 1.88, 4.02). Columns 3&4 highlight factors associated with advocacy of electronic tagging for persons with early-stage dementia: with those from urban areas less in favour when compared to their rural counterparts (OR = 0.55: 95%CI = 0.40, 0.78); and those less well-educated compared to those educated to degree-level (OR = 1.56: 1.10, 2.22 and OR = 2.18: 1.34, 3.54 for those with intermediate levels and no qualifications respectively). Finally, the patterns outlined above also broadly apply to the factors associated with those recording no opinion or mixed responses.
Despite much public health effort—and media interest—in communicating to the public that dementia progression may be amenable to modifiable risk factors [33,34,1,35] these findings confirm continuing entrenched levels of negativity amongst the general public [2–4]: with higher levels associated with better general or tacit knowledge; and more specific knowledge of risk factors mitigating somewhat against it. We note that, in this study, antipathy to dementia, which can be thought of as a precursor of stigma, is inversely associated with knowledge and contact. Moreover, respondents indicated they thought high levels of control were appropriate for people with diagnosed, even with early-stage dementia. These findings highlight the complexity of social and personal responses to dementia. Thus, knowledge about dementia and personal contact with someone who has dementia do not always engender sympathetic responses. Perhaps more in keeping with evidence from elsewhere, older age, lower education and limited knowledge are associated with perceptions of the need to exert high levels of control over people with dementia, even at early stages. Nevertheless, the findings highlighting negative characterisation of dementia and the perceived need for surveillance and control raise interrelated questions about how to address public anxiety. Are such fears related to a concern for individuals with dementia (and their social context) or do they drift into a wider social response reflecting stigmatising attitudes? Stigma is, necessarily, understood and constructed through social relationships and the rules that govern normality or what is expected . However, dementia-related stigma remains relatively untheorized .
While dementia services are exhorted to overcome what Kitwood  termed a malignant social psychology, the social exclusion of people living with dementia is still prevalent and this is likely to continue while families and communities lack resources to produce change. To what extent a sense of shame, embarrassment about dementia symptoms, fear of accidents, or difficulties in negotiating social barriers are risk factors for exclusion is unclear. Thus, while it is important to distinguish between actual and the psychological barriers to social engagement there is a likely interaction between the inherent challenges of dementia, the anticipated problems by the patient or family caregiver, and the environmental capacity to deal with them, that reinforces a ‘housebound’ exclusion. While this persists, contact with people with dementia will remain limited and public dementia anxiety high.
Strengths and limitations
This is a cross-sectional study and therefore no conclusions can be drawn about causation. While the measures and questions that tap into attitudes to dementia are not validated, it is important to note that there are no well-established and validated measures of dementia-related stigma. Moreover, the questions asked do provide sufficient breadth for the attitudes examined to be sufficiently nuanced for analysis.
In this examination of attitudes to dementia we found that respondents emphasised its challenging nature, reporting high levels of negativity across a range of issues. The study also highlighted a lack of knowledge of risk factors about dementia. Given that the underlying dementia knowledge-base is rapidly developing, with both a research focus on the modifiability of risk factors, and a shift towards a more dynamic open debate about all aspects of the disease, it seems appropriate that future attitudinal research should reflect this and focus more attention on how this cultural shift can be realized and what psychological and institutional barriers still remain within this current more fluid situation. More research is needed to establish effective strategies to increase knowledge about dementia in the lay population. Strategies to facilitate stronger social inclusion for people with dementia and their carers should be a priority for research and policy.
- 1. Kivipelto M, Hakansson K (2017). A rare success against Alzheimer’s. Scientific American, April 2017
- 2. McParland P, Devine P, Innes A, Gayle V (2012). Dementia knowledge and attitudes of the general public in Northern Ireland: an analysis of national survey data. International Psychogeriatrics 24:10, 1600–1613
- 3. Cahill S, Pierce M, Werner P, Darley A, Bobersky A. (2015). A Systematic Review of the Public’s Knowledge of Alzheimer’s Disease and Dementia. Alzheimer’s Disease and Associated Disorders, Volume 29 Issue 3
- 4. Glynn RW, Shelley E, Lawlor BA (2017). Public knowledge and understanding of dementia—evidence from a notional survey in Ireland. Age and Ageing 46: 865–869 pmid:28531240
- 5. Handley M, Bunn F, Goodman C (2015). Interventions that support the creation of dementia friendly environments in health care: protocol for a realist review Systematic Reviews 4:18
- 6. Ward R, Clark A, Campbell S, Graham B, Kullberg A, Manji K et al (2017). The lived neighborhood: understanding how people with dementia engage with their local environment. International Psychogeriatrics.
- 7. Burke D, Hickie I, Breakspear M, Gotz J. (2007). Possibilities for the prevention and treatment of cognitive impairment and dementia. British Journal of Psychiatry 190: 371–372 pmid:17470948
- 8. World Health Organization (2012). Alzheimer’s Disease International: Dementia: A public health priority. Geneva, Switzerand
- 9. Prince M (2017). Progress on dementia—leaving no one behind. Lancet Vol 390, December 16
- 10. World Health Organisation (2017). Global action plan on the public health response to dementia 2017–2025. ISBN 978–92–4–151348–7. Geneva, Switzerland
- 11. Mukadam N, Livingston G (2012). Reducing the stigma assocaited with dementia: approaches and goals. Aging Health 8(4), 377–386
- 12. Bamford SM, Holley-Moore G, Watson J, editors. (2014) New perspectives and approaches to understanding dementia and stigma: a compendium of essays. ILC-UK (International Longevity Centre: www.ilcuk.org.uk)
- 13. Rasmussen J (2014). Chapter 3: The impact of stigma for decision making, diagnosis, treatment and care of the individual, in Bamford et al “New perspectives and approaches to understanding dementia and stigma: a compendium of essays” ILC-UK
- 14. Rüsch N, Angermeyer MC, Corrigan PW (2005). Mental illness stigma: Concepts, consequences, and initiatives to reduce stigma. European Psychiatry 20(8): 529–539 pmid:16171984
- 15. Werner P, Mittelmann PH, Goldstein MA, Heinik MD (2011). Family Stigma an Caregiver Burden in Alzheimer’s Disease. The Gerontologist
- 16. Goffman E (1963). Stigma and social identity. Stigma: notes on the management of spoiled identity. Englewood Cliffs, NJ, Prentice Hall: 1–40
- 17. Scambler G (1998). Stigma and disease. The Lancet 352(9133): 1054–1055
- 18. Welsh S, Hassiotis A, O’Mahoney G, Deahl M (2003). Big brother is watching you—the ethical implications of electronic surveillance measures in the elderly with dementia and in adults with learning difficulties. Aging & Mental Health 7(5): 372–375
- 19. Hillman A, Latimer J (2017). Cultural representations of dementia. PLOS Medicine
- 20. Herrmann LK, Welter E, Leverenz J, Lerner AJ, Udelson N, Kanetsky C et al (2018). A Systematic Review of Dementia-related Stigma Research: Can We Move the Stigma Dial? American Journal of Geriatric Psychiatry, Mar;26(3):316–331. Epub 2017 Sep 12. pmid:29426607
- 21. Liu D, Hinton L, Tran C, Hinton D & Barker JC (2008). Re-examining the relationships among dementia, stigma, and aging in immigrant Chinese and Vietnamese family caregivers. Journal of Cross Cultural Gerontology, Sep; 23(3): 283–299 pmid:18665444
- 22. Hoe J, Jesnick L, Turner R, Leavey G & Livingston G (2017). Caring for relatives with agitation at home: a qualitative study of positive coping strategies. British Journal of Psychiatry Open Jan; 3(1): 34–40. Published online 2017 Feb 9. pmid:28243464
- 23. Halek M, Bartholomeyczek S (2012). Description of the behaviour of wandering in people with dementia living in nursing homes—a review of the literature. Scandinavian Journal of Caring Science 26(2): 404–413
- 24. Cipriani G, Lucetti C, Nuti A, Danti S (2014). Wandering and dementia. Psychogeriatrics 14(2): 135–142 pmid:24661471
- 25. Bharucha AJ, Anand V, Forlizzi J, Dew MA, Reynolds CF, Stevens S et al (2009). Intelligent assistive technology applications to dementia care: current capabilities, limitations, and future challenges. The American Journal of Geriatric Psychiatry 17: 88–104 pmid:18849532
- 26. NILT (2014). Northern Ireland Life and Times Survey (2014) Technical Notes. Available at (last accessed 17/05/2017): ark.ac.uk/nilt/2014/tech14.pdf
- 27. ARK website (last accessed 08/05/2018): http://www.ark.ac.uk/nilt/2014/
- 28. Devine P (2016) Discussion paper. Public attitudes and knowledge of dementia: Northern Ireland, Republic of Ireland and Scotland. www.ark.ac.uk/ageing
- 29. StataCorp (2013). Stata Statistical Software: Release 13. College Station, TX: StataCorp LP
- 30. Muthen LK, Muthen BO (2008). Mplus user’s guide, CA: Author
- 31. Smith B, Ali S, Quach H (2014). Public knowledge and beliefs about dementia risk reduction: a national survey of Australians. BMC Public Health 14:661 pmid:24972448
- 32. O’Reilly D, Rosato M, Johnston F, Catney G, Brolly M (2012). Cohort description: The Northern Ireland Longitudinal Study (NILS): International Journal of Epidemiology; 41: 634–641. pmid:21296852
- 33. Elwood P, Galante J, Pickering J, Palmer S, Bayer A, Ben-Shlomo Y et al (2013). Healthy lifestyles reduce the incidence of chronic diseases and dementia: evidence from the Caerphilly Cohort Study. PLoS ONE 8(12)
- 34. Nganda T, Lehtisalo J, Solomon A, Levalahti E, Ahtiluoto T, Antikainen R et al (2015). A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised control trial. The Lancet Vol 385: June 6
- 35. Livingstone J, Sommerlad A, Orgeta V, Costafreda SG, Huntley J, Ames D et al. (2017). Dementia Prevention, Intervention, and Care. Lancet 390: 2673–734 pmid:28735855
- 36. Link B G, Phelan JC (2001). Conceptualizing Stigma. Annual Review of Sociology 27: 363–385
- 37. Behuniak SM. (2011) The living dead? The construction of people with Alzheimer’s disease as zombies. Ageing and Society 31(1): 70–92
- 38. Kitwood T (1997). Dementia reconsidered: the person comes first. Milton Keynes, Open University Press